window.__IS_SSR__=true
window.__INITIAL_STATE__={
"attachmentsReducer": {
"audio_0": {
"type": "attachments",
"id": "audio_0",
"imgSizes": {
"kqedFullSize": {
"file": "https://ww2.kqed.org/news/wp-content/themes/KQED-unified/img/audio_bgs/background0.jpg"
}
}
},
"audio_1": {
"type": "attachments",
"id": "audio_1",
"imgSizes": {
"kqedFullSize": {
"file": "https://ww2.kqed.org/news/wp-content/themes/KQED-unified/img/audio_bgs/background1.jpg"
}
}
},
"audio_2": {
"type": "attachments",
"id": "audio_2",
"imgSizes": {
"kqedFullSize": {
"file": "https://ww2.kqed.org/news/wp-content/themes/KQED-unified/img/audio_bgs/background2.jpg"
}
}
},
"audio_3": {
"type": "attachments",
"id": "audio_3",
"imgSizes": {
"kqedFullSize": {
"file": "https://ww2.kqed.org/news/wp-content/themes/KQED-unified/img/audio_bgs/background3.jpg"
}
}
},
"audio_4": {
"type": "attachments",
"id": "audio_4",
"imgSizes": {
"kqedFullSize": {
"file": "https://ww2.kqed.org/news/wp-content/themes/KQED-unified/img/audio_bgs/background4.jpg"
}
}
},
"placeholder": {
"type": "attachments",
"id": "placeholder",
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"medium": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-800x533.jpg",
"width": 800,
"height": 533,
"mimeType": "image/jpeg"
},
"medium_large": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-768x512.jpg",
"width": 768,
"height": 512,
"mimeType": "image/jpeg"
},
"large": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1020x680.jpg",
"width": 1020,
"height": 680,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"fd-lrg": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"fd-med": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1020x680.jpg",
"width": 1020,
"height": 680,
"mimeType": "image/jpeg"
},
"fd-sm": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-800x533.jpg",
"width": 800,
"height": 533,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"xxsmall": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"xsmall": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"small": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"xlarge": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1020x680.jpg",
"width": 1020,
"height": 680,
"mimeType": "image/jpeg"
},
"full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1-1920x1280.jpg",
"width": 1920,
"height": 1280,
"mimeType": "image/jpeg"
},
"guest-author-32": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 32,
"height": 32,
"mimeType": "image/jpeg"
},
"guest-author-50": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 50,
"height": 50,
"mimeType": "image/jpeg"
},
"guest-author-64": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 64,
"height": 64,
"mimeType": "image/jpeg"
},
"guest-author-96": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 96,
"height": 96,
"mimeType": "image/jpeg"
},
"guest-author-128": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 128,
"height": 128,
"mimeType": "image/jpeg"
},
"detail": {
"file": "https://cdn.kqed.org/wp-content/uploads/2025/01/KQED-Default-Image-816638274-1333x1333-1-160x160.jpg",
"width": 160,
"height": 160,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/2024/12/KQED-Default-Image-816638274-2000x1333-1.jpg",
"width": 2000,
"height": 1333
}
}
},
"news_11989313": {
"type": "attachments",
"id": "news_11989313",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "11989313",
"found": true
},
"title": "Tech AI Illustrations",
"publishDate": 1717711326,
"status": "inherit",
"parent": 11989308,
"modified": 1787170163,
"caption": "The ChatGPT logo is displayed on an iPhone in New York, May 18, 2023. The San Francisco-based company says ChatGPT for Teens is tailored for children ages 13 to 17. Parents, educators and child development experts have been sounding alarms over children's use of AI chatbots.",
"credit": "Jaap Arriens/NurPhoto via Getty Images",
"altTag": null,
"description": null,
"imgSizes": {
"medium": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-800x533.jpg",
"width": 800,
"height": 533,
"mimeType": "image/jpeg"
},
"large": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1020x680.jpg",
"width": 1020,
"height": 680,
"mimeType": "image/jpeg"
},
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"2048x2048": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-2048x1365.jpg",
"width": 2048,
"height": 1365,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1920x1280.jpg",
"width": 1920,
"height": 1280,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-scaled-e1760733694503.jpg",
"width": 2000,
"height": 1334
}
},
"isLoading": false,
"fetchFailed": false
},
"news_11947437": {
"type": "attachments",
"id": "news_11947437",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "11947437",
"found": true
},
"parent": 11947432,
"imgSizes": {
"twentyfourteen-full-width": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-1038x576.jpg",
"width": 1038,
"mimeType": "image/jpeg",
"height": 576
},
"thumbnail": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-160x107.jpg",
"width": 160,
"mimeType": "image/jpeg",
"height": 107
},
"post-thumbnail": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-672x372.jpg",
"width": 672,
"mimeType": "image/jpeg",
"height": 372
},
"kqedFullSize": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut.jpg",
"width": 1920,
"height": 1280
},
"large": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-1020x680.jpg",
"width": 1020,
"mimeType": "image/jpeg",
"height": 680
},
"1536x1536": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-1536x1024.jpg",
"width": 1536,
"mimeType": "image/jpeg",
"height": 1024
},
"medium": {
"file": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut-800x533.jpg",
"width": 800,
"mimeType": "image/jpeg",
"height": 533
}
},
"publishDate": 1682360244,
"modified": 1682369415,
"caption": "Balboa High School teacher Michael Rosenberg checks the text from an essay assignment to see how much was generated by AI, at the school in San Francisco on April 17, 2023. The text highlighted in yellow is idenitified as AI generated text.",
"description": null,
"title": "RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut",
"credit": "Beth LaBerge/KQED",
"status": "inherit",
"altTag": "A closeup of a laptop screen with text on display.",
"isLoading": false,
"fetchFailed": false
},
"news_12094085": {
"type": "attachments",
"id": "news_12094085",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12094085",
"found": true
},
"title": "073126-Therapy-and-AI-MG-CM",
"publishDate": 1786038609,
"status": "inherit",
"parent": 12094081,
"modified": 1786038671,
"caption": "Illustration by Miguel Gutierrez Jr., CalMatters; iStock\n",
"credit": "Miguel Gutierrez Jr./CalMatters",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-160x107.jpeg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-1536x1024.jpeg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-672x372.jpeg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-1038x576.jpeg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-1200x675.jpeg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM-600x600.jpeg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/073126-Therapy-and-AI-MG-CM.jpeg",
"width": 2000,
"height": 1333
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12060375": {
"type": "attachments",
"id": "news_12060375",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12060375",
"found": true
},
"title": "US-TECH-AI-ALTMAN",
"publishDate": 1760733500,
"status": "inherit",
"parent": 12060365,
"modified": 1760733569,
"caption": "OpenAI CEO Sam Altman speaks at OpenAI DevDay, the company's annual conference for developers, in San Francisco, California, on Oct. 6, 2025. ",
"credit": "Benjamin Legendre/AFP via Getty Images",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/SamAltmanGetty-160x108.jpg",
"width": 160,
"height": 108,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/SamAltmanGetty-1536x1034.jpg",
"width": 1536,
"height": 1034,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/SamAltmanGetty-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/SamAltmanGetty-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/SamAltmanGetty.jpg",
"width": 2000,
"height": 1347
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12087570": {
"type": "attachments",
"id": "news_12087570",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12087570",
"found": true
},
"title": "CaliforniaSealCM",
"publishDate": 1781541800,
"status": "inherit",
"parent": 12087559,
"modified": 1781541837,
"caption": "Every year, California agencies report to the state technology department on their use of \"high-risk\" decision systems. The seal of the State of California on the Edmund G. Brown Jr. State Office Building in San Francisco on Jan. 28, 2022. ",
"credit": "Martin do Nascimento for CalMatters",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CaliforniaSealCM.jpg",
"width": 2000,
"height": 1333
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12084126": {
"type": "attachments",
"id": "news_12084126",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12084126",
"found": true
},
"title": "260518-MUSK-ALTMAN-TRIAL-VB-06-KQED",
"publishDate": 1779135074,
"status": "inherit",
"parent": 0,
"modified": 1779136158,
"caption": "The verdict is read in the trial in which Elon Musk claimed that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 18, 2026.",
"credit": "Vicki Behringer for KQED",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-160x90.jpg",
"width": 160,
"height": 90,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-1536x864.jpg",
"width": 1536,
"height": 864,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-06-KQED.jpg",
"width": 2000,
"height": 1125
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12083803": {
"type": "attachments",
"id": "news_12083803",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12083803",
"found": true
},
"title": "Elon Musk v. OpenAI Trial Continues In California",
"publishDate": 1778804683,
"status": "inherit",
"parent": 12083612,
"modified": 1778805456,
"caption": "OpenAI's attorney William Savitt makes remarks during a press conference outside the courthouse at the Ronald V. Dellums Federal Building on May 6, 2026, in Oakland, California. Elon Musk invested in OpenAI early on, believing it would be a non-profit, but is now suing OpenAI and its CEO, Sam Altman, for allegedly deceiving him by developing OpenAI into a for-profit company. ",
"credit": "Benjamin Fanjoy/Getty Images",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty.jpg",
"width": 2000,
"height": 1333
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12083392": {
"type": "attachments",
"id": "news_12083392",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12083392",
"found": true
},
"title": "260512-MUSK-ALTMAN-TRIAL-VB-03-KQED",
"publishDate": 1778625867,
"status": "inherit",
"parent": 0,
"modified": 1778625974,
"caption": "Open AI CEO Sam Altman testifies as a video of him is played on a screen in the trial in which Elon Musk claims that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 12, 2026.",
"credit": "Vicki Behringer for KQED",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-160x90.jpg",
"width": 160,
"height": 90,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-1536x864.jpg",
"width": 1536,
"height": 864,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED.jpg",
"width": 2000,
"height": 1125
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12083235": {
"type": "attachments",
"id": "news_12083235",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12083235",
"found": true
},
"title": "Elon Musk v. OpenAI Trial Continues In California",
"publishDate": 1778545779,
"status": "inherit",
"parent": 12083224,
"modified": 1778546322,
"caption": "OpenAI CEO Sam Altman arrives at the Ronald V. Dellums Federal Building on April 30, 2026, in Oakland, California. Elon Musk invested in OpenAI early on, believing it would be a nonprofit, but is now suing OpenAI and its CEO, Sam Altman, for allegedly deceiving him by developing OpenAI into a for-profit company. ",
"credit": "Benjamin Fanjoy/Getty Images",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-160x107.jpg",
"width": 160,
"height": 107,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-1536x1024.jpg",
"width": 1536,
"height": 1024,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty.jpg",
"width": 2000,
"height": 1333
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12082344": {
"type": "attachments",
"id": "news_12082344",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12082344",
"found": true
},
"title": "260504-MUSK-ALTMAN-VB-04-KQED",
"publishDate": 1777936709,
"status": "inherit",
"parent": 0,
"modified": 1777936770,
"caption": "Open AI President Greg Brockman testifies in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit in Oakland on May 4, 2026.",
"credit": "Vicki Behringer for KQED",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-160x90.jpg",
"width": 160,
"height": 90,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg",
"width": 1536,
"height": 864,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-04-KQED.jpg",
"width": 2000,
"height": 1125
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12082068": {
"type": "attachments",
"id": "news_12082068",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12082068",
"found": true
},
"title": "CANADA-SHOOTING-CRIME",
"publishDate": 1777656025,
"status": "inherit",
"parent": 12082064,
"modified": 1777677641,
"caption": "A young boy brings flowers to a memorial in honor of the victims of one of Canada's deadliest shootings in Tumbler Ridge, British Columbia, Canada, on Feb. 13, 2026. An 18-year-old carried out a mass shooting in a remote mining town, killing six people at a local school, after slaying her mother and stepbrother. Canadian Police Commander Dwayne McDonald said authorities still don't know the motive in the Feb. 10 mass shooting, but the shooter, who took her own life, was known to have mental health issues. ",
"credit": "Paige Taylor White/AFP via Getty Images",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-160x110.jpg",
"width": 160,
"height": 110,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-1536x1053.jpg",
"width": 1536,
"height": 1053,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty.jpg",
"width": 2000,
"height": 1371
}
},
"isLoading": false,
"fetchFailed": false
},
"news_12081681": {
"type": "attachments",
"id": "news_12081681",
"meta": {
"index": "attachments_1716263798",
"site": "news",
"id": "12081681",
"found": true
},
"title": "260428-MUSK-ALTMAN-VB-03-KQED-1",
"publishDate": 1777416108,
"status": "inherit",
"parent": 12081603,
"modified": 1777508469,
"caption": "Representing Microsoft, Russell Coan (left) speaks as Elon Musk watches in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026.",
"credit": "Vicki Behringer for KQED",
"altTag": null,
"description": null,
"imgSizes": {
"thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-160x90.jpg",
"width": 160,
"height": 90,
"mimeType": "image/jpeg"
},
"1536x1536": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-1536x864.jpg",
"width": 1536,
"height": 864,
"mimeType": "image/jpeg"
},
"post-thumbnail": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-672x372.jpg",
"width": 672,
"height": 372,
"mimeType": "image/jpeg"
},
"twentyfourteen-full-width": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-1038x576.jpg",
"width": 1038,
"height": 576,
"mimeType": "image/jpeg"
},
"npr-cds-wide": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-1200x675.jpg",
"width": 1200,
"height": 675,
"mimeType": "image/jpeg"
},
"npr-cds-square": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1-600x600.jpg",
"width": 600,
"height": 600,
"mimeType": "image/jpeg"
},
"kqedFullSize": {
"file": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-03-KQED-1.jpg",
"width": 2000,
"height": 1125
}
},
"isLoading": false,
"fetchFailed": false
}
},
"audioPlayerReducer": {
"postId": "stream_live",
"isPaused": true,
"isPlaying": false,
"pfsActive": false,
"pledgeModalIsOpen": true,
"playerDrawerIsOpen": false,
"liveAudioPlayStartedAt": 0,
"liveAudioPlayContext": ""
},
"authorsReducer": {
"byline_news_12096019": {
"type": "authors",
"id": "byline_news_12096019",
"meta": {
"override": true
},
"slug": "byline_news_12096019",
"name": "Barbara Ortutay and Kaitlyn Huamani, AP Technology ",
"isLoading": false
},
"byline_news_12094081": {
"type": "authors",
"id": "byline_news_12094081",
"meta": {
"override": true
},
"slug": "byline_news_12094081",
"name": "Ana B. Ibarra, CalMatters",
"isLoading": false
},
"byline_news_12087559": {
"type": "authors",
"id": "byline_news_12087559",
"meta": {
"override": true
},
"slug": "byline_news_12087559",
"name": "Khari Johnson, CalMatters",
"isLoading": false
},
"byline_news_12082064": {
"type": "authors",
"id": "byline_news_12082064",
"meta": {
"override": true
},
"slug": "byline_news_12082064",
"name": "Matt O’Brien, Associated Press, and Nisa Khan, KQED",
"isLoading": false
},
"rachael-myrow": {
"type": "authors",
"id": "251",
"meta": {
"index": "authors_1716337520",
"id": "251",
"found": true
},
"name": "Rachael Myrow",
"firstName": "Rachael",
"lastName": "Myrow",
"slug": "rachael-myrow",
"email": "rmyrow@kqed.org",
"display_author_email": true,
"staff_mastheads": [
"news"
],
"title": "Senior Editor of KQED's Silicon Valley News Desk",
"bio": "• I write and edit stories about how Silicon Valley power and policies shape everyday life in California. I’m also passionate about making Bay Area history and culture more accessible to a broad public. • I’ve been a journalist for most of my life, starting in high school with The Franklin Press in Los Angeles, where I grew up. While earning my first degree in English at UC Berkeley, I got my start in public radio at KALX-FM. After completing a second degree in journalism at Cal, I landed my first professional job at Marketplace, then moved on to KPCC (now LAist), and then KQED, where I hosted The California Report for more than seven years. • My reporting has appeared on NPR, The World, WBUR’s \u003ci>Here & Now\u003c/i>, and the BBC. I also guest host for KQED’s \u003ci>Forum\u003c/i>, as well as the Commonwealth Club in San Francisco. • I speak periodically on media, democracy and technology issues, and do voiceover work for documentaries and educational video projects. • Outside of the studio, you'll find me hiking Bay Area trails and whipping up Insta-ready meals in my kitchen. • I do not accept gifts, money, or favors from anyone connected to my reporting, I don't pay people for information, and I do not support or donate to political causes. • I strive to treat the people I report on with fairness, honesty, and respect. I also recognize there are often multiple sides to a story and work to verify information through multiple sources and documentation. If I get something wrong, I correct it.",
"avatar": "https://secure.gravatar.com/avatar/3ed8a3c5f26025e4a805ba634cda8cd764906b97438a6d9ce216b044e340da4e?s=600&d=blank&r=g",
"twitter": "rachaelmyrow",
"bluesky": null,
"facebook": null,
"instagram": null,
"linkedin": "https://www.linkedin.com/in/rachaelmyrow/",
"sites": [
{
"site": "arts",
"roles": [
"administrator"
]
},
{
"site": "news",
"roles": [
"edit_others_posts",
"editor"
]
},
{
"site": "futureofyou",
"roles": [
"editor"
]
},
{
"site": "bayareabites",
"roles": [
"editor"
]
},
{
"site": "stateofhealth",
"roles": [
"editor"
]
},
{
"site": "science",
"roles": [
"editor"
]
},
{
"site": "food",
"roles": [
"editor"
]
},
{
"site": "forum",
"roles": [
"editor"
]
},
{
"site": "liveblog",
"roles": [
"author"
]
}
],
"headData": {
"title": "Rachael Myrow | KQED",
"description": "Senior Editor of KQED's Silicon Valley News Desk",
"ogImgSrc": "https://secure.gravatar.com/avatar/3ed8a3c5f26025e4a805ba634cda8cd764906b97438a6d9ce216b044e340da4e?s=600&d=blank&r=g",
"twImgSrc": "https://secure.gravatar.com/avatar/3ed8a3c5f26025e4a805ba634cda8cd764906b97438a6d9ce216b044e340da4e?s=600&d=blank&r=g"
},
"isLoading": false,
"link": "/author/rachael-myrow"
},
"ecruzguevarra": {
"type": "authors",
"id": "8654",
"meta": {
"index": "authors_1716337520",
"id": "8654",
"found": true
},
"name": "Ericka Cruz Guevarra",
"firstName": "Ericka",
"lastName": "Cruz Guevarra",
"slug": "ecruzguevarra",
"email": "ecruzguevarra@kqed.org",
"display_author_email": true,
"staff_mastheads": [
"news"
],
"title": "Producer, The Bay Podcast",
"bio": "Ericka Cruz Guevarra is host of \u003ca href=\"https://www.kqed.org/podcasts/thebay\">\u003cem>The Bay\u003c/em>\u003c/a> podcast at KQED. Before host, she was the show’s producer. Her work in that capacity includes a three-part reported series on policing in Vallejo, which won a 2020 excellence in journalism award from the Society of Professional Journalists. Ericka has worked as a breaking news reporter at Oregon Public Broadcasting, helped produce the Code Switch podcast, and was KQED’s inaugural Raul Ramirez Diversity Fund intern. She’s also an alumna of NPR’s Next Generation Radio program. Send her an email if you have strong feelings about whether Fairfield and Suisun City are the Bay. Ericka is represented by SAG-AFTRA.",
"avatar": "https://secure.gravatar.com/avatar/accefc1de4bafc7e234b69bcb9b6f6536c130fd84709bc3069ef11a78787e1bb?s=600&d=blank&r=g",
"twitter": "NotoriousECG",
"bluesky": null,
"facebook": null,
"instagram": null,
"linkedin": null,
"sites": [
{
"site": "arts",
"roles": [
"subscriber"
]
},
{
"site": "news",
"roles": [
"editor",
"manage_categories"
]
},
{
"site": "futureofyou",
"roles": [
"subscriber"
]
},
{
"site": "stateofhealth",
"roles": [
"subscriber"
]
},
{
"site": "science",
"roles": [
"editor"
]
},
{
"site": "forum",
"roles": [
"subscriber"
]
}
],
"headData": {
"title": "Ericka Cruz Guevarra | KQED",
"description": "Producer, The Bay Podcast",
"ogImgSrc": "https://secure.gravatar.com/avatar/accefc1de4bafc7e234b69bcb9b6f6536c130fd84709bc3069ef11a78787e1bb?s=600&d=blank&r=g",
"twImgSrc": "https://secure.gravatar.com/avatar/accefc1de4bafc7e234b69bcb9b6f6536c130fd84709bc3069ef11a78787e1bb?s=600&d=blank&r=g"
},
"isLoading": false,
"link": "/author/ecruzguevarra"
},
"amontecillo": {
"type": "authors",
"id": "11649",
"meta": {
"index": "authors_1716337520",
"id": "11649",
"found": true
},
"name": "Alan Montecillo",
"firstName": "Alan",
"lastName": "Montecillo",
"slug": "amontecillo",
"email": "amontecillo@kqed.org",
"display_author_email": false,
"staff_mastheads": [
"news"
],
"title": "KQED Contributor",
"bio": "Alan Montecillo is the senior editor of \u003cem>\u003ca href=\"http://kqed.org/thebay\">The Bay\u003c/a>, \u003c/em> KQED's local news podcast. Before moving to the Bay Area, he worked as a senior talk show producer for WILL in Champaign-Urbana, Illinois and at Oregon Public Broadcasting in Portland, Oregon. He has won journalism awards from the Society of Professional Journalists Northern California, the Public Media Journalists Association, The Signal Awards, and has also received a regional Edward R. Murrow award. Alan is a Filipino American from Hong Kong and a graduate of Reed College.",
"avatar": "https://secure.gravatar.com/avatar/268f53a2066899ea9efdd8a5900218b17c822d9862cbc9326cccec53ef41d349?s=600&d=blank&r=g",
"twitter": "alanmontecillo",
"bluesky": null,
"facebook": null,
"instagram": null,
"linkedin": null,
"sites": [
{
"site": "",
"roles": [
"editor"
]
},
{
"site": "news",
"roles": [
"editor",
"manage_categories"
]
}
],
"headData": {
"title": "Alan Montecillo | KQED",
"description": "KQED Contributor",
"ogImgSrc": "https://secure.gravatar.com/avatar/268f53a2066899ea9efdd8a5900218b17c822d9862cbc9326cccec53ef41d349?s=600&d=blank&r=g",
"twImgSrc": "https://secure.gravatar.com/avatar/268f53a2066899ea9efdd8a5900218b17c822d9862cbc9326cccec53ef41d349?s=600&d=blank&r=g"
},
"isLoading": false,
"link": "/author/amontecillo"
},
"jessicakariisa": {
"type": "authors",
"id": "11831",
"meta": {
"index": "authors_1716337520",
"id": "11831",
"found": true
},
"name": "Jessica Kariisa",
"firstName": "Jessica",
"lastName": "Kariisa",
"slug": "jessicakariisa",
"email": "jkariisa@kqed.org",
"display_author_email": false,
"staff_mastheads": [],
"title": "Producer, The Bay",
"bio": "Jessica Kariisa is the producer of The Bay. She first joined KQED as an intern for The California Report Magazine, after which she became an on-call producer. She reported a Bay Curious episode on the use of rap lyrics in criminal trials which won a Society of Professional Journalists award in 2023 for Excellence in Features Journalism and the 2023 Signal Award for Best Conversation Starter. She’s worked on podcasts for Snap Judgment and American Public Media. Before embarking on her audio career, she was a music journalist.\r\n\r\nJessica Kariisa is represented by SAG-AFTRA.",
"avatar": "https://secure.gravatar.com/avatar/522e97b28ac0b3d411621c0fde8f3419c7b71ea2c5c262c57c4bb600963f328c?s=600&d=blank&r=g",
"twitter": null,
"bluesky": null,
"facebook": null,
"instagram": null,
"linkedin": null,
"sites": [
{
"site": "arts",
"roles": [
"author"
]
},
{
"site": "news",
"roles": [
"editor",
"manage_categories"
]
}
],
"headData": {
"title": "Jessica Kariisa | KQED",
"description": "Producer, The Bay",
"ogImgSrc": "https://secure.gravatar.com/avatar/522e97b28ac0b3d411621c0fde8f3419c7b71ea2c5c262c57c4bb600963f328c?s=600&d=blank&r=g",
"twImgSrc": "https://secure.gravatar.com/avatar/522e97b28ac0b3d411621c0fde8f3419c7b71ea2c5c262c57c4bb600963f328c?s=600&d=blank&r=g"
},
"isLoading": false,
"link": "/author/jessicakariisa"
},
"kdebenedetti": {
"type": "authors",
"id": "11913",
"meta": {
"index": "authors_1716337520",
"id": "11913",
"found": true
},
"name": "Katie DeBenedetti",
"firstName": "Katie",
"lastName": "DeBenedetti",
"slug": "kdebenedetti",
"email": "kdebenedetti@kqed.org",
"display_author_email": false,
"staff_mastheads": [
"news",
"science"
],
"title": "KQED Contributor",
"bio": "Katie DeBenedetti is a digital reporter covering daily news for the Express Desk. Prior to joining KQED as a culture reporting intern in January 2024, she covered education and city government for the Napa Valley Register.",
"avatar": "https://secure.gravatar.com/avatar/6e31073cb8f7e4214ab03f42771d0f45?s=600&d=blank&r=g",
"twitter": null,
"facebook": null,
"instagram": null,
"linkedin": null,
"sites": [
{
"site": "news",
"roles": [
"author"
]
},
{
"site": "science",
"roles": [
"author"
]
},
{
"site": "liveblog",
"roles": [
"author"
]
}
],
"headData": {
"title": "Katie DeBenedetti | KQED",
"description": "KQED Contributor",
"ogImgSrc": "https://secure.gravatar.com/avatar/6e31073cb8f7e4214ab03f42771d0f45?s=600&d=blank&r=g",
"twImgSrc": "https://secure.gravatar.com/avatar/6e31073cb8f7e4214ab03f42771d0f45?s=600&d=blank&r=g"
},
"isLoading": false,
"link": "/author/kdebenedetti"
}
},
"pagesReducer": {
"news_tag_chatgpt": {
"type": "terms",
"id": "news_32668",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "32668",
"score": 9.893801
},
"featImg": null,
"name": "ChatGPT",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "ChatGPT Archives | KQED News",
"ogDescription": null
},
"ttid": 32685,
"slug": "chatgpt",
"isLoading": false,
"title": "ChatGPT",
"pageMeta": {
"site": "news",
"WpPageTemplate": "page-topic-editorial",
"currentPage": 1
},
"blocks": [
{
"blockName": "kqed/post-list",
"attrs": {
"layout": "cardArticle2",
"query": "posts/news?tag=chatgpt",
"seeMore": false,
"paginated": true,
"page": 1
}
},
{
"blockName": "kqed/ad"
}
]
}
},
"pfsSessionReducer": {},
"postsReducer": {
"stream_live": {
"type": "live",
"id": "stream_live",
"audioUrl": "https://streams.kqed.org/kqedradio",
"title": "Live Stream",
"excerpt": "Live Stream information currently unavailable.",
"link": "/radio",
"featImg": "",
"label": {
"name": "KQED Live",
"link": "/"
}
},
"stream_kqedNewscast": {
"type": "posts",
"id": "stream_kqedNewscast",
"audioUrl": "https://www.kqed.org/.stream/anon/radio/RDnews/newscast.mp3?_=1",
"title": "KQED Newscast",
"featImg": "",
"label": {
"name": "88.5 FM",
"link": "/"
}
},
"news_12096019": {
"type": "posts",
"id": "news_12096019",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12096019",
"score": null,
"sort": [
1787174767000
]
},
"guestAuthors": [],
"slug": "openai-introduces-chatgpt-for-teens-promising-a-more-age-appropriate-chatbot",
"title": "OpenAI Introduces ChatGPT for Teens, Promising a More Age-Appropriate Chatbot",
"publishDate": 1787174767,
"format": "standard",
"headTitle": "OpenAI Introduces ChatGPT for Teens, Promising a More Age-Appropriate Chatbot | KQED",
"labelTerm": {},
"content": "\u003cp> OpenAI on Tuesday introduced a version of ChatGPT designed for teenagers — the first generation to grow up \u003ca href=\"https://www.kqed.org/news/tag/artificial-intelligence\">with artificial intelligence\u003c/a> — who are already using it for schoolwork, questions about daily life and even companionship.\u003c/p>\n\n\n\n\u003cp>The San Francisco-based company says ChatGPT for Teens is tailored for kids ages 13 to 17 with stronger protections including content restrictions around things like suicide, self-harm and romantic or sexual chats. It also provides homework and study support designed to help students learn rather than \u003ca href=\"https://www.kqed.org/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning\">spit out answers\u003c/a> and school essays.\u003c/p>\n\n\n\n\u003cp>The idea is to guide teens toward healthy AI use in an age-appropriate environment, the company said.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg\" alt=\"\" class=\"wp-image-12094819\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-1536x1025.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Three teenage boys look at their smartphone screens on April 11, 2026. In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media. (Matt Cardy/Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“We want to treat teens like teens, which means that we have to make sure that we’re showing up with the right developmental stage when we’re not either talking down to them or treating them like kids, but we’re also making sure that they’re not exposed to material that they shouldn’t be exposed to,” said Ann O’Leary, vice president of global policy at OpenAI.\u003c/p>\n\n\n\n\u003cp>The effectiveness of the safeguards remains to be seen, said Brendan Bouffard, associate general counsel for AI and privacy at children’s advocacy group Fairplay. He said it is unclear whether OpenAI’s actions are enough to address young users’ compulsive use of and emotional dependence on chatbots.\u003c/p>\n\n\n\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>“Announcements like these can be seen as a step in the right direction,” he said. “But I also would caution against parents being lured into this false sense of safety based on the announcement.”\u003c/p>\n\n\n\n\u003ch2 class=\"wp-block-heading\">Risks abound in widespread teen usage of AI chatbots\u003c/h2>\n\n\n\n\u003cp>Parents, educators and child development experts have been sounding alarms over children’s use of AI chatbots, which have been blamed for facilitating cheating on schoolwork and even \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">suicide\u003c/a>. While even adults can fall victim to anthropomorphizing AI and developing unhealthy relationships with it, teenagers’ brains are not yet fully developed and they can be particularly vulnerable.\u003c/p>\n\n\n\n\u003cp>Last year, research from a watchdog group found that \u003ca href=\"https://www.kqed.org/news/12048270/teens-under-the-influence-of-chatbots\">ChatGPT would tell 13-year-olds\u003c/a> how to get drunk and high, instruct them on how to conceal eating disorders and even compose a heartbreaking suicide letter to their parents if asked. \u003c/p>\n\n\n\n\u003cp>In interactions with researchers posing as vulnerable teens, ChatGPT typically warned against risky activity but went on to deliver startlingly detailed and personalized plans for drug use, calorie-restricted diets or self-injury.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg\" alt=\"\" class=\"wp-image-11998856\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Several bills addressing generative artificial intelligence are moving through the state Legislature in California’s piecemeal approach to regulation. (Jaque Silva/SOPA Images/LightRocket via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media, a group that studies and advocates for using digital media sensibly.\u003c/p>\n\n\n\n\u003cp>OpenAI CEO Sam Altman has said the company is trying to study “emotional overreliance” on the technology, describing it last year as a “really common thing” with young people.\u003c/p>\n\n\n\n\u003cp>For users of ChatGPT for Teens, the chatbot is prevented from suggesting it has personal feelings toward the user or implying that it is conscious or experiences emotions, according to OpenAI. That’s in addition to blocking romantic or sexual chats.\u003c/p>\n\n\n\n\u003cp>“We went through and identified what are the hypothetical cues that a model could give that might make a teenager kind of develop a relationship to it,” said Allison Mishkin, head of child development at OpenAI.\u003c/p>\n\n\n\n\u003ch2 class=\"wp-block-heading\">A new version of the chatbot expands safety protections\u003c/h2>\n\n\n\n\u003cp>OpenAI doesn’t verify users’ ages, but it already uses age assurance to estimate if someone is under 18 based on factors such as their types of queries. \u003c/p>\n\n\n\n\u003cp>If someone is identified or identifies themselves as a minor, they are automatically placed into the teen version of the chatbot. \u003c/p>\n\n\n\n\u003cp>This is similar to Meta’s approach to \u003ca href=\"https://apnews.com/article/instagram-meta-teen-accounts-kids-425a05cc37f81b6ff2b7439d3884c54a\">teen accounts on Instagram\u003c/a>, which have stricter content, chat and privacy restrictions than regular accounts.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. (Beata Zawrzel/NurPhoto via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>To use parental controls, both the teen user and their parent or guardian have to opt in. But O’Leary said the idea with the teen chatbot is to “make sure that this is safe, even if you don’t use parental controls.” \u003c/p>\n\n\n\n\u003cp>Parents with linked teen accounts can set “quiet hours” when their teen can’t access ChatGPT, and receive safety notifications in limited high-risk situations, such as the possibility of a user hurting themself.\u003c/p>\n\n\n\n\u003cp>“We are adding additional notifications related to eating disorders, while limiting what is shared and focusing on moments when offline support may matter most,” OpenAI said.\u003c/p>\n\n\n\n\n\n\u003cp>Bouffard of Fairplay said some of what he would consider to be the most important safeguards, such as restricting ChatGPT’s long-term memory, are not in place by default in the teen accounts and would require the intervention of parental controls. \u003c/p>\n\n\n\n\u003cp>That creates “a major gap in safety” despite the default protections around content and other areas of concern, he said.\u003c/p>\n\n\n\n\u003cp>For homework help, OpenAI said the teen chatbot is designed to guide students to come up with answers on their own. \u003c/p>\n\n\n\n\u003cp>The company already offers a version of ChatGPT for teachers, and tailoring a model to help kids with studying and homework could give OpenAI more ways to bring its product to schools.\u003c/p>\n\n\n\n\u003cp>\u003c/p>\n\u003cp>“We continue to also invest in expanding interactive learning because research shows that people learn more effectively when they actively engage and struggle with concepts,” Mishkin said.\u003c/p>\n\n",
"blocks": [
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp> OpenAI on Tuesday introduced a version of ChatGPT designed for teenagers — the first generation to grow up \u003ca href=\"https://www.kqed.org/news/tag/artificial-intelligence\">with artificial intelligence\u003c/a> — who are already using it for schoolwork, questions about daily life and even companionship.\u003c/p>\n",
"innerContent": [
"\n\u003cp> OpenAI on Tuesday introduced a version of ChatGPT designed for teenagers — the first generation to grow up \u003ca href=\"https://www.kqed.org/news/tag/artificial-intelligence\">with artificial intelligence\u003c/a> — who are already using it for schoolwork, questions about daily life and even companionship.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The San Francisco-based company says ChatGPT for Teens is tailored for kids ages 13 to 17 with stronger protections including content restrictions around things like suicide, self-harm and romantic or sexual chats. It also provides homework and study support designed to help students learn rather than \u003ca href=\"https://www.kqed.org/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning\">spit out answers\u003c/a> and school essays.\u003c/p>\n",
"innerContent": [
"\n\u003cp>The San Francisco-based company says ChatGPT for Teens is tailored for kids ages 13 to 17 with stronger protections including content restrictions around things like suicide, self-harm and romantic or sexual chats. It also provides homework and study support designed to help students learn rather than \u003ca href=\"https://www.kqed.org/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning\">spit out answers\u003c/a> and school essays.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The idea is to guide teens toward healthy AI use in an age-appropriate environment, the company said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>The idea is to guide teens toward healthy AI use in an age-appropriate environment, the company said.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12094819,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Matt Cardy/Getty Images",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-1536x1025.jpg 1536w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg\" alt=\"\" class=\"wp-image-12094819\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-1536x1025.jpg 1536w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\"/>\u003cfigcaption class=\"wp-element-caption\">Three teenage boys look at their smartphone screens on April 11, 2026. In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media.\u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg\" alt=\"\" class=\"wp-image-12094819\"/>\u003cfigcaption class=\"wp-element-caption\">Three teenage boys look at their smartphone screens on April 11, 2026. In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media.\u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“We want to treat teens like teens, which means that we have to make sure that we’re showing up with the right developmental stage when we’re not either talking down to them or treating them like kids, but we’re also making sure that they’re not exposed to material that they shouldn’t be exposed to,” said Ann O’Leary, vice president of global policy at OpenAI.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“We want to treat teens like teens, which means that we have to make sure that we’re showing up with the right developmental stage when we’re not either talking down to them or treating them like kids, but we’re also making sure that they’re not exposed to material that they shouldn’t be exposed to,” said Ann O’Leary, vice president of global policy at OpenAI.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The effectiveness of the safeguards remains to be seen, said Brendan Bouffard, associate general counsel for AI and privacy at children’s advocacy group Fairplay. He said it is unclear whether OpenAI’s actions are enough to address young users’ compulsive use of and emotional dependence on chatbots.\u003c/p>\n",
"innerContent": [
"\n\u003cp>The effectiveness of the safeguards remains to be seen, said Brendan Bouffard, associate general counsel for AI and privacy at children’s advocacy group Fairplay. He said it is unclear whether OpenAI’s actions are enough to address young users’ compulsive use of and emotional dependence on chatbots.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "fullwidth"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“Announcements like these can be seen as a step in the right direction,” he said. “But I also would caution against parents being lured into this false sense of safety based on the announcement.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“Announcements like these can be seen as a step in the right direction,” he said. “But I also would caution against parents being lured into this false sense of safety based on the announcement.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/heading",
"attrs": {
"text": "Risks abound in widespread teen usage of AI chatbots",
"level": 2
},
"innerHTML": "\n\u003ch2 class=\"wp-block-heading\">Risks abound in widespread teen usage of AI chatbots\u003c/h2>\n",
"innerContent": [
"\n\u003ch2 class=\"wp-block-heading\">Risks abound in widespread teen usage of AI chatbots\u003c/h2>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Parents, educators and child development experts have been sounding alarms over children’s use of AI chatbots, which have been blamed for facilitating cheating on schoolwork and even \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">suicide\u003c/a>. While even adults can fall victim to anthropomorphizing AI and developing unhealthy relationships with it, teenagers’ brains are not yet fully developed and they can be particularly vulnerable.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Parents, educators and child development experts have been sounding alarms over children’s use of AI chatbots, which have been blamed for facilitating cheating on schoolwork and even \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">suicide\u003c/a>. While even adults can fall victim to anthropomorphizing AI and developing unhealthy relationships with it, teenagers’ brains are not yet fully developed and they can be particularly vulnerable.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Last year, research from a watchdog group found that \u003ca href=\"https://www.kqed.org/news/12048270/teens-under-the-influence-of-chatbots\">ChatGPT would tell 13-year-olds\u003c/a> how to get drunk and high, instruct them on how to conceal eating disorders and even compose a heartbreaking suicide letter to their parents if asked. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Last year, research from a watchdog group found that \u003ca href=\"https://www.kqed.org/news/12048270/teens-under-the-influence-of-chatbots\">ChatGPT would tell 13-year-olds\u003c/a> how to get drunk and high, instruct them on how to conceal eating disorders and even compose a heartbreaking suicide letter to their parents if asked. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>In interactions with researchers posing as vulnerable teens, ChatGPT typically warned against risky activity but went on to deliver startlingly detailed and personalized plans for drug use, calorie-restricted diets or self-injury.\u003c/p>\n",
"innerContent": [
"\n\u003cp>In interactions with researchers posing as vulnerable teens, ChatGPT typically warned against risky activity but went on to deliver startlingly detailed and personalized plans for drug use, calorie-restricted diets or self-injury.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 11998856,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Jaque Silva/SOPA Images/LightRocket via Getty Images",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1920x1280.jpg 1920w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg\" alt=\"\" class=\"wp-image-11998856\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1920x1280.jpg 1920w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\"/>\u003cfigcaption class=\"wp-element-caption\">Several bills addressing generative artificial intelligence are moving through the state Legislature in California’s piecemeal approach to regulation.\u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg\" alt=\"\" class=\"wp-image-11998856\"/>\u003cfigcaption class=\"wp-element-caption\">Several bills addressing generative artificial intelligence are moving through the state Legislature in California’s piecemeal approach to regulation.\u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media, a group that studies and advocates for using digital media sensibly.\u003c/p>\n",
"innerContent": [
"\n\u003cp>In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media, a group that studies and advocates for using digital media sensibly.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>OpenAI CEO Sam Altman has said the company is trying to study “emotional overreliance” on the technology, describing it last year as a “really common thing” with young people.\u003c/p>\n",
"innerContent": [
"\n\u003cp>OpenAI CEO Sam Altman has said the company is trying to study “emotional overreliance” on the technology, describing it last year as a “really common thing” with young people.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>For users of ChatGPT for Teens, the chatbot is prevented from suggesting it has personal feelings toward the user or implying that it is conscious or experiences emotions, according to OpenAI. That’s in addition to blocking romantic or sexual chats.\u003c/p>\n",
"innerContent": [
"\n\u003cp>For users of ChatGPT for Teens, the chatbot is prevented from suggesting it has personal feelings toward the user or implying that it is conscious or experiences emotions, according to OpenAI. That’s in addition to blocking romantic or sexual chats.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“We went through and identified what are the hypothetical cues that a model could give that might make a teenager kind of develop a relationship to it,” said Allison Mishkin, head of child development at OpenAI.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“We went through and identified what are the hypothetical cues that a model could give that might make a teenager kind of develop a relationship to it,” said Allison Mishkin, head of child development at OpenAI.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/heading",
"attrs": {
"text": "A new version of the chatbot expands safety protections",
"level": 2
},
"innerHTML": "\n\u003ch2 class=\"wp-block-heading\">A new version of the chatbot expands safety protections\u003c/h2>\n",
"innerContent": [
"\n\u003ch2 class=\"wp-block-heading\">A new version of the chatbot expands safety protections\u003c/h2>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>OpenAI doesn’t verify users’ ages, but it already uses age assurance to estimate if someone is under 18 based on factors such as their types of queries. \u003c/p>\n",
"innerContent": [
"\n\u003cp>OpenAI doesn’t verify users’ ages, but it already uses age assurance to estimate if someone is under 18 based on factors such as their types of queries. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>If someone is identified or identifies themselves as a minor, they are automatically placed into the teen version of the chatbot. \u003c/p>\n",
"innerContent": [
"\n\u003cp>If someone is identified or identifies themselves as a minor, they are automatically placed into the teen version of the chatbot. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>This is similar to Meta’s approach to \u003ca href=\"https://apnews.com/article/instagram-meta-teen-accounts-kids-425a05cc37f81b6ff2b7439d3884c54a\">teen accounts on Instagram\u003c/a>, which have stricter content, chat and privacy restrictions than regular accounts.\u003c/p>\n",
"innerContent": [
"\n\u003cp>This is similar to Meta’s approach to \u003ca href=\"https://apnews.com/article/instagram-meta-teen-accounts-kids-425a05cc37f81b6ff2b7439d3884c54a\">teen accounts on Instagram\u003c/a>, which have stricter content, chat and privacy restrictions than regular accounts.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12036125,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Beata Zawrzel/NurPhoto via Getty Images",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\"/>\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\"/>\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>To use parental controls, both the teen user and their parent or guardian have to opt in. But O’Leary said the idea with the teen chatbot is to “make sure that this is safe, even if you don’t use parental controls.” \u003c/p>\n",
"innerContent": [
"\n\u003cp>To use parental controls, both the teen user and their parent or guardian have to opt in. But O’Leary said the idea with the teen chatbot is to “make sure that this is safe, even if you don’t use parental controls.” \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Parents with linked teen accounts can set “quiet hours” when their teen can’t access ChatGPT, and receive safety notifications in limited high-risk situations, such as the possibility of a user hurting themself.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Parents with linked teen accounts can set “quiet hours” when their teen can’t access ChatGPT, and receive safety notifications in limited high-risk situations, such as the possibility of a user hurting themself.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“We are adding additional notifications related to eating disorders, while limiting what is shared and focusing on moments when offline support may matter most,” OpenAI said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“We are adding additional notifications related to eating disorders, while limiting what is shared and focusing on moments when offline support may matter most,” OpenAI said.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed-v2/related-articles",
"attrs": {
"postIds": [
"news_12095554"
],
"label": "Related Articles",
"maxPosts": 1,
"showImages": false,
"heroImage": "https://ww2.kqed.org/app/uploads/sites/10/2023/04/RS64554_005_KQED_ChatGPTBalboaHS_04172023-qut.jpg",
"query": "",
"labelWasCleared": false,
"heroImageAlt": "",
"heroImageLink": "",
"labelLink": "",
"alignment": "right",
"target": "_self",
"manualLinks": [],
"posts": []
},
"innerHTML": "",
"innerContent": [],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Bouffard of Fairplay said some of what he would consider to be the most important safeguards, such as restricting ChatGPT’s long-term memory, are not in place by default in the teen accounts and would require the intervention of parental controls. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Bouffard of Fairplay said some of what he would consider to be the most important safeguards, such as restricting ChatGPT’s long-term memory, are not in place by default in the teen accounts and would require the intervention of parental controls. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>That creates “a major gap in safety” despite the default protections around content and other areas of concern, he said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>That creates “a major gap in safety” despite the default protections around content and other areas of concern, he said.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>For homework help, OpenAI said the teen chatbot is designed to guide students to come up with answers on their own. \u003c/p>\n",
"innerContent": [
"\n\u003cp>For homework help, OpenAI said the teen chatbot is designed to guide students to come up with answers on their own. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The company already offers a version of ChatGPT for teachers, and tailoring a model to help kids with studying and homework could give OpenAI more ways to bring its product to schools.\u003c/p>\n",
"innerContent": [
"\n\u003cp>The company already offers a version of ChatGPT for teachers, and tailoring a model to help kids with studying and homework could give OpenAI more ways to bring its product to schools.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "floatright"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“We continue to also invest in expanding interactive learning because research shows that people learn more effectively when they actively engage and struggle with concepts,” Mishkin said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“We continue to also invest in expanding interactive learning because research shows that people learn more effectively when they actively engage and struggle with concepts,” Mishkin said.\u003c/p>\n"
],
"innerBlocks": []
}
],
"excerpt": "The idea is to guide teens toward healthy AI use in an age-appropriate environment, the San Francisco-based company said.",
"status": "publish",
"parent": 0,
"modified": 1787174775,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 25,
"wordCount": 1030
},
"headData": {
"title": "OpenAI Introduces ChatGPT for Teens, Promising a More Age-Appropriate Chatbot | KQED",
"description": "The idea is to guide teens toward healthy AI use in an age-appropriate environment, the San Francisco-based company said.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "OpenAI Introduces ChatGPT for Teens, Promising a More Age-Appropriate Chatbot",
"datePublished": "2026-08-19T14:26:07-07:00",
"dateModified": "2026-08-19T14:26:15-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"source": "The Associated Press",
"sticky": false,
"nprByline": "Barbara Ortutay and Kaitlyn Huamani, AP Technology ",
"nprStoryId": "kqed-12096019",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"showOnAuthorArchivePages": "No",
"articleAge": "0",
"path": "/news/12096019/openai-introduces-chatgpt-for-teens-promising-a-more-age-appropriate-chatbot",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp> OpenAI on Tuesday introduced a version of ChatGPT designed for teenagers — the first generation to grow up \u003ca href=\"https://www.kqed.org/news/tag/artificial-intelligence\">with artificial intelligence\u003c/a> — who are already using it for schoolwork, questions about daily life and even companionship.\u003c/p>\n\n\n\n\u003cp>The San Francisco-based company says ChatGPT for Teens is tailored for kids ages 13 to 17 with stronger protections including content restrictions around things like suicide, self-harm and romantic or sexual chats. It also provides homework and study support designed to help students learn rather than \u003ca href=\"https://www.kqed.org/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning\">spit out answers\u003c/a> and school essays.\u003c/p>\n\n\n\n\u003cp>The idea is to guide teens toward healthy AI use in an age-appropriate environment, the company said.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg\" alt=\"\" class=\"wp-image-12094819\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/MetaTeensGetty-1536x1025.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Three teenage boys look at their smartphone screens on April 11, 2026. In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media. (Matt Cardy/Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“We want to treat teens like teens, which means that we have to make sure that we’re showing up with the right developmental stage when we’re not either talking down to them or treating them like kids, but we’re also making sure that they’re not exposed to material that they shouldn’t be exposed to,” said Ann O’Leary, vice president of global policy at OpenAI.\u003c/p>\n\n\n\n\u003cp>The effectiveness of the safeguards remains to be seen, said Brendan Bouffard, associate general counsel for AI and privacy at children’s advocacy group Fairplay. He said it is unclear whether OpenAI’s actions are enough to address young users’ compulsive use of and emotional dependence on chatbots.\u003c/p>\n\n\n\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>“Announcements like these can be seen as a step in the right direction,” he said. “But I also would caution against parents being lured into this false sense of safety based on the announcement.”\u003c/p>\n\n\n\n\u003ch2 class=\"wp-block-heading\">Risks abound in widespread teen usage of AI chatbots\u003c/h2>\n\n\n\n\u003cp>Parents, educators and child development experts have been sounding alarms over children’s use of AI chatbots, which have been blamed for facilitating cheating on schoolwork and even \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">suicide\u003c/a>. While even adults can fall victim to anthropomorphizing AI and developing unhealthy relationships with it, teenagers’ brains are not yet fully developed and they can be particularly vulnerable.\u003c/p>\n\n\n\n\u003cp>Last year, research from a watchdog group found that \u003ca href=\"https://www.kqed.org/news/12048270/teens-under-the-influence-of-chatbots\">ChatGPT would tell 13-year-olds\u003c/a> how to get drunk and high, instruct them on how to conceal eating disorders and even compose a heartbreaking suicide letter to their parents if asked. \u003c/p>\n\n\n\n\u003cp>In interactions with researchers posing as vulnerable teens, ChatGPT typically warned against risky activity but went on to deliver startlingly detailed and personalized plans for drug use, calorie-restricted diets or self-injury.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg\" alt=\"\" class=\"wp-image-11998856\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/08/ChatGPTGetty-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Several bills addressing generative artificial intelligence are moving through the state Legislature in California’s piecemeal approach to regulation. (Jaque Silva/SOPA Images/LightRocket via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In the U.S., more than 70% of teens are turning to \u003ca href=\"https://www.kqed.org/news/12038154/kids-talking-ai-companion-chatbots-stanford-researchers-say-thats-bad-idea\">AI chatbots for companionship\u003c/a> and half use AI companions regularly, according to a 2025 study from Common Sense Media, a group that studies and advocates for using digital media sensibly.\u003c/p>\n\n\n\n\u003cp>OpenAI CEO Sam Altman has said the company is trying to study “emotional overreliance” on the technology, describing it last year as a “really common thing” with young people.\u003c/p>\n\n\n\n\u003cp>For users of ChatGPT for Teens, the chatbot is prevented from suggesting it has personal feelings toward the user or implying that it is conscious or experiences emotions, according to OpenAI. That’s in addition to blocking romantic or sexual chats.\u003c/p>\n\n\n\n\u003cp>“We went through and identified what are the hypothetical cues that a model could give that might make a teenager kind of develop a relationship to it,” said Allison Mishkin, head of child development at OpenAI.\u003c/p>\n\n\n\n\u003ch2 class=\"wp-block-heading\">A new version of the chatbot expands safety protections\u003c/h2>\n\n\n\n\u003cp>OpenAI doesn’t verify users’ ages, but it already uses age assurance to estimate if someone is under 18 based on factors such as their types of queries. \u003c/p>\n\n\n\n\u003cp>If someone is identified or identifies themselves as a minor, they are automatically placed into the teen version of the chatbot. \u003c/p>\n\n\n\n\u003cp>This is similar to Meta’s approach to \u003ca href=\"https://apnews.com/article/instagram-meta-teen-accounts-kids-425a05cc37f81b6ff2b7439d3884c54a\">teen accounts on Instagram\u003c/a>, which have stricter content, chat and privacy restrictions than regular accounts.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. (Beata Zawrzel/NurPhoto via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>To use parental controls, both the teen user and their parent or guardian have to opt in. But O’Leary said the idea with the teen chatbot is to “make sure that this is safe, even if you don’t use parental controls.” \u003c/p>\n\n\n\n\u003cp>Parents with linked teen accounts can set “quiet hours” when their teen can’t access ChatGPT, and receive safety notifications in limited high-risk situations, such as the possibility of a user hurting themself.\u003c/p>\n\n\n\n\u003cp>“We are adding additional notifications related to eating disorders, while limiting what is shared and focusing on moments when offline support may matter most,” OpenAI said.\u003c/p>\n\n\n\n\n\n\u003cp>Bouffard of Fairplay said some of what he would consider to be the most important safeguards, such as restricting ChatGPT’s long-term memory, are not in place by default in the teen accounts and would require the intervention of parental controls. \u003c/p>\n\n\n\n\u003cp>That creates “a major gap in safety” despite the default protections around content and other areas of concern, he said.\u003c/p>\n\n\n\n\u003cp>For homework help, OpenAI said the teen chatbot is designed to guide students to come up with answers on their own. \u003c/p>\n\n\n\n\u003cp>The company already offers a version of ChatGPT for teachers, and tailoring a model to help kids with studying and homework could give OpenAI more ways to bring its product to schools.\u003c/p>\n\n\n\n\u003cp>\u003c/p>\n\u003cp>“We continue to also invest in expanding interactive learning because research shows that people learn more effectively when they actively engage and struggle with concepts,” Mishkin said.\u003c/p>\n\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12096019/openai-introduces-chatgpt-for-teens-promising-a-more-age-appropriate-chatbot",
"authors": [
"byline_news_12096019"
],
"categories": [
"news_31795",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_18538",
"news_36279",
"news_32668",
"news_17996",
"news_33542",
"news_35836",
"news_38",
"news_35940"
],
"featImg": "news_11989313",
"label": "source_news_12096019"
},
"news_12095554": {
"type": "posts",
"id": "news_12095554",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12095554",
"score": null,
"sort": [
1787050800000
]
},
"guestAuthors": [],
"slug": "most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning",
"title": "Most Teens Use AI for Homework, Even if Some Think They’re Missing Out on Learning",
"publishDate": 1787050800,
"format": "standard",
"headTitle": "Most Teens Use AI for Homework, Even if Some Think They’re Missing Out on Learning | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>Seven in ten teenagers are \u003ca href=\"https://www.kqed.org/news/category/technology\">using artificial intelligence\u003c/a> for their schoolwork, according to new data released by San Francisco-based nonprofit Common Sense Media. \u003c/p>\n\n\n\n\u003cp>As schools and parents grapple with how to better control students’ technology and AI access, the survey of more than 1,000 teens across the U.S. found that the vast majority are already relying on the tools, while few are learning how to use them safely. \u003c/p>\n\n\n\n\u003cp>“AI has become routine in how teens do homework,” said Michael Robb, the head of research at Common Sense Media. “I actually think the more interesting finding is kind of how they are using AI.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\">\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. (Smith Collection/Gado via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“Our biggest concern … is that even if you think you are using AI in a way that’s beneficial to your learning, you might actually be subtly undermining your learning,” he said. \u003c/p>\n\n\n\n\u003cp>The survey found that 63% of teens using AI for schoolwork are getting an answer from the tool in some form. What they do with these responses varies: a quarter said they use the AI-generated text as is, while about 31% rewrite it in their voice and just 35% say they improve it with their own knowledge. \u003c/p>\n\n\n\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>Robb said the concern with all of those uses is whether they are disrupting the learning process. The Common Sense Media report found that among the AI users surveyed, 38% say having the tools available leads them to develop fewer of their own ideas.\u003c/p>\n\n\n\n\n\n\u003cp>Two-thirds of teens said that AI helped them understand their schoolwork, but almost four in 10 reported feeling like they were missing out on learning when they used AI to complete assignments.\u003c/p>\n\n\n\n\u003cp>Schools, Robb said, should be “trying to design assignments so that AI is not removing the important thinking parts, things like drafting or revising, coming up with arguments.”\u003c/p>\n\n\n\n\u003cp>Teens reported that they believe the skills that will matter most for their future are social skills, fact-checking, critical thinking and understanding complex information. Robb said AI companies, school districts and employers should be considering if the tools available to their students are counterintuitive to those goals.\u003c/p>\n\n\n\n\u003cp>“The question is, if we’re introducing AI in all these various forms, are those things actually supporting human connection or critical thinking? Or are they replacing or disrupting them?” he asked. “Because if it’s the latter, then … we are undercutting the very things we set to value.”\u003c/p>\n\n\n\n\u003cp>While Robb said using AI to check work or get feedback can be useful, he worries about students who said they use it to brainstorm for assignments. \u003c/p>\n\n\n\n\u003cp>“They are starting with an answer before a student is trying the problem,” Robb said. “There are a couple of skills I think are at stake. Starting hard things, generating your own ideas flexibly — these are executive function skills that are still developing.”\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1335\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg\" alt=\"\" class=\"wp-image-11992395\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-800x534.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1020x681.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1536x1025.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1920x1282.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Students walk down a hallway at Fremont High School in Oakland on Oct. 10, 2023. (Laure Andrillon for CalMatters)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“This is happening at a period in adolescence where you really want to hone and strengthen these skills, and I’m concerned that we are undermining them,” he continued. \u003c/p>\n\n\n\n\u003cp>Sixty-eight percent of students who use AI said they use one of those multi-purpose tools, like ChatGPT, for their schoolwork, while four in 10 reported that they use a reading or writing assistant, like Grammarly. Smaller percentages use AI-backed research tools, image and video generators or tutors.\u003c/p>\n\n\n\n\u003cp>San Francisco’s school district in 2025 introduced an AI-powered tutoring program called Amira, which it said provides students with individualized reading support, without the need for one-on-one human tutors. Amira’s \u003ca href=\"https://amiralearning.com/research\">website\u003c/a> said studies show that the software “outperforms high dosage human tutoring at scale,” though critics have \u003ca href=\"https://sfeducation.substack.com/p/no-amiras-ai-tutor-is-not-ready-to\">disputed\u003c/a> this claim. According to district documents, SFUSD uses the program for elementary students in class and at home.\u003c/p>\n\n\n\n\u003cp>“I think [something] that schools should be looking at is making sure that AI is not replacing learning from people,” Robb said. “People throw around the word ‘personalized’ a lot … but that should really just mean that kids need to be better supported, not leaving kids along with an AI system instead of a teacher or a classmate.”\u003c/p>\n\n\n\n\u003cp>The report comes amid a parent-led movement across some major school districts to restrict school technology use, as the majority have \u003ca href=\"https://www.cosn.org/edtech-topics/state-of-edtech-leadership/\">1:1\u003c/a> device programs that give students access to a laptop or tablet both in their classrooms and at home. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\">\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. (Michael Dwyer/Associated Press)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>Many of these devices make it easy for students to access artificial intelligence technology — from the automatically generated AI summary that tops Google search result pages to access to tools like ChatGPT, Grok and Gemini — leading the national parent group \u003ca href=\"https://www.schoolsbeyondscreens.com/reform-proposals\">Schools Beyond Screens\u003c/a> to call for a moratorium on AI in schools. \u003c/p>\n\n\n\n\u003cp>The Common Sense Media survey found that while many students have access to this tech, they aren’t all discussing how to use it safely or to support their learning. \u003c/p>\n\n\n\n\u003cp>“It’s important that schools make their rules really clear,” Robb said. \u003c/p>\n\n\n\n\u003cp>Common Sense Media has also created a new curriculum that focuses on safety concerns like deepfakes and social engineering, a technique cybercriminals use to get people to reveal sensitive data or grant access to restricted systems.\u003c/p>\n\n\n\n\u003cp>\u003c/p>\n\u003cp>\u003cem>KQED’s Jose Velazquez contributed to this report.\u003c/em>\u003c/p>\n\n",
"blocks": [
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Seven in ten teenagers are \u003ca href=\"https://www.kqed.org/news/category/technology\">using artificial intelligence\u003c/a> for their schoolwork, according to new data released by San Francisco-based nonprofit Common Sense Media. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Seven in ten teenagers are \u003ca href=\"https://www.kqed.org/news/category/technology\">using artificial intelligence\u003c/a> for their schoolwork, according to new data released by San Francisco-based nonprofit Common Sense Media. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>As schools and parents grapple with how to better control students’ technology and AI access, the survey of more than 1,000 teens across the U.S. found that the vast majority are already relying on the tools, while few are learning how to use them safely. \u003c/p>\n",
"innerContent": [
"\n\u003cp>As schools and parents grapple with how to better control students’ technology and AI access, the survey of more than 1,000 teens across the U.S. found that the vast majority are already relying on the tools, while few are learning how to use them safely. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“AI has become routine in how teens do homework,” said Michael Robb, the head of research at Common Sense Media. “I actually think the more interesting finding is kind of how they are using AI.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“AI has become routine in how teens do homework,” said Michael Robb, the head of research at Common Sense Media. “I actually think the more interesting finding is kind of how they are using AI.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12058035,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Smith Collection/Gado via Getty Images",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\" />\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\" />\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“Our biggest concern … is that even if you think you are using AI in a way that’s beneficial to your learning, you might actually be subtly undermining your learning,” he said. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“Our biggest concern … is that even if you think you are using AI in a way that’s beneficial to your learning, you might actually be subtly undermining your learning,” he said. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The survey found that 63% of teens using AI for schoolwork are getting an answer from the tool in some form. What they do with these responses varies: a quarter said they use the AI-generated text as is, while about 31% rewrite it in their voice and just 35% say they improve it with their own knowledge. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The survey found that 63% of teens using AI for schoolwork are getting an answer from the tool in some form. What they do with these responses varies: a quarter said they use the AI-generated text as is, while about 31% rewrite it in their voice and just 35% say they improve it with their own knowledge. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "fullwidth"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Robb said the concern with all of those uses is whether they are disrupting the learning process. The Common Sense Media report found that among the AI users surveyed, 38% say having the tools available leads them to develop fewer of their own ideas.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Robb said the concern with all of those uses is whether they are disrupting the learning process. The Common Sense Media report found that among the AI users surveyed, 38% say having the tools available leads them to develop fewer of their own ideas.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed-v2/related-articles",
"attrs": {
"postIds": [
"news_12091964"
],
"label": "Related Articles",
"maxPosts": 1,
"showImages": false,
"heroImage": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/AIGetty.jpg",
"query": "",
"labelWasCleared": false,
"heroImageAlt": "",
"heroImageLink": "",
"labelLink": "",
"alignment": "right",
"target": "_self",
"manualLinks": [],
"posts": []
},
"innerHTML": "",
"innerContent": [],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Two-thirds of teens said that AI helped them understand their schoolwork, but almost four in 10 reported feeling like they were missing out on learning when they used AI to complete assignments.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Two-thirds of teens said that AI helped them understand their schoolwork, but almost four in 10 reported feeling like they were missing out on learning when they used AI to complete assignments.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Schools, Robb said, should be “trying to design assignments so that AI is not removing the important thinking parts, things like drafting or revising, coming up with arguments.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>Schools, Robb said, should be “trying to design assignments so that AI is not removing the important thinking parts, things like drafting or revising, coming up with arguments.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Teens reported that they believe the skills that will matter most for their future are social skills, fact-checking, critical thinking and understanding complex information. Robb said AI companies, school districts and employers should be considering if the tools available to their students are counterintuitive to those goals.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Teens reported that they believe the skills that will matter most for their future are social skills, fact-checking, critical thinking and understanding complex information. Robb said AI companies, school districts and employers should be considering if the tools available to their students are counterintuitive to those goals.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“The question is, if we’re introducing AI in all these various forms, are those things actually supporting human connection or critical thinking? Or are they replacing or disrupting them?” he asked. “Because if it’s the latter, then … we are undercutting the very things we set to value.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“The question is, if we’re introducing AI in all these various forms, are those things actually supporting human connection or critical thinking? Or are they replacing or disrupting them?” he asked. “Because if it’s the latter, then … we are undercutting the very things we set to value.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>While Robb said using AI to check work or get feedback can be useful, he worries about students who said they use it to brainstorm for assignments. \u003c/p>\n",
"innerContent": [
"\n\u003cp>While Robb said using AI to check work or get feedback can be useful, he worries about students who said they use it to brainstorm for assignments. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“They are starting with an answer before a student is trying the problem,” Robb said. “There are a couple of skills I think are at stake. Starting hard things, generating your own ideas flexibly — these are executive function skills that are still developing.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“They are starting with an answer before a student is trying the problem,” Robb said. “There are a couple of skills I think are at stake. Starting hard things, generating your own ideas flexibly — these are executive function skills that are still developing.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 11992395,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Laure Andrillon for CalMatters",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-800x534.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1020x681.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1536x1025.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1920x1282.jpg 1920w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg\" alt=\"\" class=\"wp-image-11992395\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-800x534.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1020x681.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1536x1025.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1920x1282.jpg 1920w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\" />\u003cfigcaption class=\"wp-element-caption\">Students walk down a hallway at Fremont High School in Oakland on Oct. 10, 2023. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg\" alt=\"\" class=\"wp-image-11992395\" />\u003cfigcaption class=\"wp-element-caption\">Students walk down a hallway at Fremont High School in Oakland on Oct. 10, 2023. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“This is happening at a period in adolescence where you really want to hone and strengthen these skills, and I’m concerned that we are undermining them,” he continued. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“This is happening at a period in adolescence where you really want to hone and strengthen these skills, and I’m concerned that we are undermining them,” he continued. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Sixty-eight percent of students who use AI said they use one of those multi-purpose tools, like ChatGPT, for their schoolwork, while four in 10 reported that they use a reading or writing assistant, like Grammarly. Smaller percentages use AI-backed research tools, image and video generators or tutors.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Sixty-eight percent of students who use AI said they use one of those multi-purpose tools, like ChatGPT, for their schoolwork, while four in 10 reported that they use a reading or writing assistant, like Grammarly. Smaller percentages use AI-backed research tools, image and video generators or tutors.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>San Francisco’s school district in 2025 introduced an AI-powered tutoring program called Amira, which it said provides students with individualized reading support, without the need for one-on-one human tutors. Amira’s \u003ca href=\"https://amiralearning.com/research\">website\u003c/a> said studies show that the software “outperforms high dosage human tutoring at scale,” though critics have \u003ca href=\"https://sfeducation.substack.com/p/no-amiras-ai-tutor-is-not-ready-to\">disputed\u003c/a> this claim. According to district documents, SFUSD uses the program for elementary students in class and at home.\u003c/p>\n",
"innerContent": [
"\n\u003cp>San Francisco’s school district in 2025 introduced an AI-powered tutoring program called Amira, which it said provides students with individualized reading support, without the need for one-on-one human tutors. Amira’s \u003ca href=\"https://amiralearning.com/research\">website\u003c/a> said studies show that the software “outperforms high dosage human tutoring at scale,” though critics have \u003ca href=\"https://sfeducation.substack.com/p/no-amiras-ai-tutor-is-not-ready-to\">disputed\u003c/a> this claim. According to district documents, SFUSD uses the program for elementary students in class and at home.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“I think [something] that schools should be looking at is making sure that AI is not replacing learning from people,” Robb said. “People throw around the word ‘personalized’ a lot … but that should really just mean that kids need to be better supported, not leaving kids along with an AI system instead of a teacher or a classmate.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“I think [something] that schools should be looking at is making sure that AI is not replacing learning from people,” Robb said. “People throw around the word ‘personalized’ a lot … but that should really just mean that kids need to be better supported, not leaving kids along with an AI system instead of a teacher or a classmate.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The report comes amid a parent-led movement across some major school districts to restrict school technology use, as the majority have \u003ca href=\"https://www.cosn.org/edtech-topics/state-of-edtech-leadership/\">1:1\u003c/a> device programs that give students access to a laptop or tablet both in their classrooms and at home. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The report comes amid a parent-led movement across some major school districts to restrict school technology use, as the majority have \u003ca href=\"https://www.cosn.org/edtech-topics/state-of-edtech-leadership/\">1:1\u003c/a> device programs that give students access to a laptop or tablet both in their classrooms and at home. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 11985952,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Michael Dwyer/Associated Press",
"alt": ""
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\" />\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\" />\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Many of these devices make it easy for students to access artificial intelligence technology — from the automatically generated AI summary that tops Google search result pages to access to tools like ChatGPT, Grok and Gemini — leading the national parent group \u003ca href=\"https://www.schoolsbeyondscreens.com/reform-proposals\">Schools Beyond Screens\u003c/a> to call for a moratorium on AI in schools. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Many of these devices make it easy for students to access artificial intelligence technology — from the automatically generated AI summary that tops Google search result pages to access to tools like ChatGPT, Grok and Gemini — leading the national parent group \u003ca href=\"https://www.schoolsbeyondscreens.com/reform-proposals\">Schools Beyond Screens\u003c/a> to call for a moratorium on AI in schools. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>The Common Sense Media survey found that while many students have access to this tech, they aren’t all discussing how to use it safely or to support their learning. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The Common Sense Media survey found that while many students have access to this tech, they aren’t all discussing how to use it safely or to support their learning. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>“It’s important that schools make their rules really clear,” Robb said. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“It’s important that schools make their rules really clear,” Robb said. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>Common Sense Media has also created a new curriculum that focuses on safety concerns like deepfakes and social engineering, a technique cybercriminals use to get people to reveal sensitive data or grant access to restricted systems.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Common Sense Media has also created a new curriculum that focuses on safety concerns like deepfakes and social engineering, a technique cybercriminals use to get people to reveal sensitive data or grant access to restricted systems.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "floatright"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": {
"dropCap": false
},
"innerHTML": "\n\u003cp>\u003cem>KQED’s Jose Velazquez contributed to this report.\u003c/em>\u003c/p>\n",
"innerContent": [
"\n\u003cp>\u003cem>KQED’s Jose Velazquez contributed to this report.\u003c/em>\u003c/p>\n"
],
"innerBlocks": []
}
],
"excerpt": "A new study from Bay Area-based Common Sense Media suggests AI tools could be disrupting students’ development of key skills, like critical thinking and executive function.",
"status": "publish",
"parent": 0,
"modified": 1787009179,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 24,
"wordCount": 992
},
"headData": {
"title": "Most Teens Use AI for Homework, Even if Some Think They’re Missing Out on Learning | KQED",
"description": "A new study from Bay Area-based Common Sense Media suggests AI tools could be disrupting students’ development of key skills, like critical thinking and executive function.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Most Teens Use AI for Homework, Even if Some Think They’re Missing Out on Learning",
"datePublished": "2026-08-18T04:00:00-07:00",
"dateModified": "2026-08-17T16:26:19-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 18540,
"slug": "education",
"name": "Education"
},
"sticky": false,
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>Seven in ten teenagers are \u003ca href=\"https://www.kqed.org/news/category/technology\">using artificial intelligence\u003c/a> for their schoolwork, according to new data released by San Francisco-based nonprofit Common Sense Media. \u003c/p>\n\n\n\n\u003cp>As schools and parents grapple with how to better control students’ technology and AI access, the survey of more than 1,000 teens across the U.S. found that the vast majority are already relying on the tools, while few are learning how to use them safely. \u003c/p>\n\n\n\n\u003cp>“AI has become routine in how teens do homework,” said Michael Robb, the head of research at Common Sense Media. “I actually think the more interesting finding is kind of how they are using AI.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\">\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. (Smith Collection/Gado via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“Our biggest concern … is that even if you think you are using AI in a way that’s beneficial to your learning, you might actually be subtly undermining your learning,” he said. \u003c/p>\n\n\n\n\u003cp>The survey found that 63% of teens using AI for schoolwork are getting an answer from the tool in some form. What they do with these responses varies: a quarter said they use the AI-generated text as is, while about 31% rewrite it in their voice and just 35% say they improve it with their own knowledge. \u003c/p>\n\n\n\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>Robb said the concern with all of those uses is whether they are disrupting the learning process. The Common Sense Media report found that among the AI users surveyed, 38% say having the tools available leads them to develop fewer of their own ideas.\u003c/p>\n\n\n\n\n\n\u003cp>Two-thirds of teens said that AI helped them understand their schoolwork, but almost four in 10 reported feeling like they were missing out on learning when they used AI to complete assignments.\u003c/p>\n\n\n\n\u003cp>Schools, Robb said, should be “trying to design assignments so that AI is not removing the important thinking parts, things like drafting or revising, coming up with arguments.”\u003c/p>\n\n\n\n\u003cp>Teens reported that they believe the skills that will matter most for their future are social skills, fact-checking, critical thinking and understanding complex information. Robb said AI companies, school districts and employers should be considering if the tools available to their students are counterintuitive to those goals.\u003c/p>\n\n\n\n\u003cp>“The question is, if we’re introducing AI in all these various forms, are those things actually supporting human connection or critical thinking? Or are they replacing or disrupting them?” he asked. “Because if it’s the latter, then … we are undercutting the very things we set to value.”\u003c/p>\n\n\n\n\u003cp>While Robb said using AI to check work or get feedback can be useful, he worries about students who said they use it to brainstorm for assignments. \u003c/p>\n\n\n\n\u003cp>“They are starting with an answer before a student is trying the problem,” Robb said. “There are a couple of skills I think are at stake. Starting hard things, generating your own ideas flexibly — these are executive function skills that are still developing.”\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1335\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg\" alt=\"\" class=\"wp-image-11992395\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-800x534.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1020x681.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1536x1025.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/101023-AI-College-Toby-Reed-LA-CM-21-copy-1920x1282.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">Students walk down a hallway at Fremont High School in Oakland on Oct. 10, 2023. (Laure Andrillon for CalMatters)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“This is happening at a period in adolescence where you really want to hone and strengthen these skills, and I’m concerned that we are undermining them,” he continued. \u003c/p>\n\n\n\n\u003cp>Sixty-eight percent of students who use AI said they use one of those multi-purpose tools, like ChatGPT, for their schoolwork, while four in 10 reported that they use a reading or writing assistant, like Grammarly. Smaller percentages use AI-backed research tools, image and video generators or tutors.\u003c/p>\n\n\n\n\u003cp>San Francisco’s school district in 2025 introduced an AI-powered tutoring program called Amira, which it said provides students with individualized reading support, without the need for one-on-one human tutors. Amira’s \u003ca href=\"https://amiralearning.com/research\">website\u003c/a> said studies show that the software “outperforms high dosage human tutoring at scale,” though critics have \u003ca href=\"https://sfeducation.substack.com/p/no-amiras-ai-tutor-is-not-ready-to\">disputed\u003c/a> this claim. According to district documents, SFUSD uses the program for elementary students in class and at home.\u003c/p>\n\n\n\n\u003cp>“I think [something] that schools should be looking at is making sure that AI is not replacing learning from people,” Robb said. “People throw around the word ‘personalized’ a lot … but that should really just mean that kids need to be better supported, not leaving kids along with an AI system instead of a teacher or a classmate.”\u003c/p>\n\n\n\n\u003cp>The report comes amid a parent-led movement across some major school districts to restrict school technology use, as the majority have \u003ca href=\"https://www.cosn.org/edtech-topics/state-of-edtech-leadership/\">1:1\u003c/a> device programs that give students access to a laptop or tablet both in their classrooms and at home. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\">\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. (Michael Dwyer/Associated Press)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>Many of these devices make it easy for students to access artificial intelligence technology — from the automatically generated AI summary that tops Google search result pages to access to tools like ChatGPT, Grok and Gemini — leading the national parent group \u003ca href=\"https://www.schoolsbeyondscreens.com/reform-proposals\">Schools Beyond Screens\u003c/a> to call for a moratorium on AI in schools. \u003c/p>\n\n\n\n\u003cp>The Common Sense Media survey found that while many students have access to this tech, they aren’t all discussing how to use it safely or to support their learning. \u003c/p>\n\n\n\n\u003cp>“It’s important that schools make their rules really clear,” Robb said. \u003c/p>\n\n\n\n\u003cp>Common Sense Media has also created a new curriculum that focuses on safety concerns like deepfakes and social engineering, a technique cybercriminals use to get people to reveal sensitive data or grant access to restricted systems.\u003c/p>\n\n\n\n\u003cp>\u003c/p>\n\u003cp>\u003cem>KQED’s Jose Velazquez contributed to this report.\u003c/em>\u003c/p>\n\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12095554/most-teens-use-ai-for-homework-even-if-some-think-theyre-missing-out-on-learning",
"authors": [
"11913"
],
"categories": [
"news_18540",
"news_8",
"news_248"
],
"tags": [
"news_25184",
"news_32668",
"news_20013",
"news_27626",
"news_22602",
"news_3457",
"news_348",
"news_1631",
"news_20385",
"news_98"
],
"featImg": "news_11947437",
"label": "news"
},
"news_12094081": {
"type": "posts",
"id": "news_12094081",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12094081",
"score": null,
"sort": [
1786038933000
]
},
"guestAuthors": [],
"slug": "as-ai-therapists-dish-out-advice-california-lawmakers-try-to-set-some-limits",
"title": "As AI ‘Therapists’ Dish Out Advice, California Lawmakers Try to Set Some Limits",
"publishDate": 1786038933,
"format": "standard",
"headTitle": "As AI ‘Therapists’ Dish Out Advice, California Lawmakers Try to Set Some Limits | KQED",
"labelTerm": {},
"content": "\u003cp>\u003c!-- Creative Commons Attribution-NonCommercial-NoDerivatives https://creativecommons.org/licenses/by-nc-nd/4.0/ -->\u003c/p>\n\u003cp>\u003cem>This story was originally published by \u003ca href=\"https://calmatters.org/\">CalMatters\u003c/a>. \u003ca href=\"https://calmatters.org/subscribe-to-calmatters/\">Sign up\u003c/a> for their newsletters.\u003c/em>\u003c/p>\n\u003cp>In the matter of a few clicks and keystrokes, anyone can find themselves deep in conversation with “Psychologist,” \u003ca href=\"https://www.kqed.org/news/tag/ai\">a chatbot\u003c/a> character that describes itself as an expert in “empathy” and “active listening.”\u003c/p>\n\u003cp>It responds to messages of anxiety and sadness with reassurance. It asks questions and offers advice; it even responds with italicized nonverbal cues: “\u003cem>The psychologist’s expression softens with compassion\u003c/em>.” In smaller font, a disclaimer reads: “This is A.I. and not a real person. Treat everything it says as fiction.”\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>As more people turn to chatbots for mental health help and advice — including by some estimates \u003ca href=\"https://sph.brown.edu/news/2025-11-18/teens-ai-chatbots\">one in eight teens and young adults\u003c/a> — legislators are in a race to keep up with the quickly developing technology.\u003c/p>\n\u003cp>Among the latest efforts is a bill that would add restrictions around the use of chatbot companions and other artificial intelligence-powered tools that people use to support mental healthcare or in lieu of professional help.\u003c/p>\n\u003cp>Sen. \u003ca href=\"https://calmatters.digitaldemocracy.org/legislators/steve-padilla-165435\">Steve Padilla\u003c/a>, a San Diego Democrat and the bill’s author, said the intent is to draw a line: AI can be used for administrative purposes and to support clinicians, but it cannot practice psychotherapy.\u003c/p>\n\u003cfigure id=\"attachment_12094084\" class=\"wp-caption aligncenter\" style=\"max-width: 1980px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12094084\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472.jpg\" alt=\"\" width=\"1980\" height=\"1231\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472-160x99.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472-1536x955.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-caption-text\">Senator Steve Padilla, commissioner on the California Coastal Commission. \u003ccite>(Nelvin C. Cepeda/The San Diego Union-Tribune via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“We’re dealing with a new impactful technology that is unfolding and is deployed in our world at light speed, and so it is both extremely powerful and consequential,” Padilla said.\u003c/p>\n\u003cp>The debate has taken on new urgency amid wrongful death lawsuits — some playing out in California federal courts — that accuse AI chatbot makers of contributing to users’ suicides.\u003c/p>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202520260sb903\">Senate Bill 903\u003c/a> would ban companies from advertising chatbots as therapy. It would prohibit AI from making therapeutic decisions without the review of a licensed professional and require health providers to disclose and get a patient’s permission before using AI tools to record therapy sessions or to triage mental healthcare.\u003c/p>\n\u003ch2>Consumer protections vs. innovation\u003c/h2>\n\u003cp>The bill is backed by professional associations that represent psychologists, therapists and counselors. Also behind the bill is the National Union of Healthcare Workers, which has been leading a push against unchecked use of artificial intelligence and digital tools in mental healthcare.\u003c/p>\n\u003cp>Most recently, the union \u003ca href=\"https://calmatters.org/health/2026/07/kaiser-mental-health-ai-complaint/\">filed a complaint\u003c/a> with state regulators against Kaiser Permanente for its alleged use of an automated algorithm in mental health triaging.\u003c/p>\n\u003cp>Padilla and supporters argue that the legislation is needed to protect consumers from AI in both formal and informal spaces, but also to protect licensed professionals from being displaced by AI.\u003c/p>\n\u003cfigure id=\"attachment_11963409\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11963409\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED.jpg\" alt='A large modern building with the words \"Kaiser Permanente\" across the top.' width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">The Kaiser Permanente Oakland Medical Center in Oakland on Oct. 4, 2023. \u003ccite>(Martin do Nascimento/KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/hearings/279268#t=669&f=e564faf4d88bf6c955f241d3fb6627e1\">Le Ondra Clark Harvey\u003c/a>, chief executive officer at the California Behavioral Health Association, told lawmakers that without clear safeguards, chatbots can provide inaccurate information or mishandle crisis situations. “The difference between a licensed clinician and an automated response is not technical. It can be life altering,” Harvey said.\u003c/p>\n\u003cp>Opposition to the bill has largely focused around the restrictions of AI use in formal clinical spaces. TechNet, which represents tech-sector companies, argues the legislation could restrict beneficial uses of AI in healthcare and stifle innovation. \u003ca href=\"https://calmatters.digitaldemocracy.org/hearings/279268#t=900&f=e564faf4d88bf6c955f241d3fb6627e1\">Robert Boykin\u003c/a>, TechNet’s executive director for California, said the bill could limit the use of tools that help expand access to mental healthcare.\u003c/p>\n\u003cp>“At a time when every county in California faces a shortage of behavioral healthcare workers, SB 903 still puts a clinician bottleneck in front of the intake and screening tools that help patients reach care faster,” Boykin said in an emailed statement.\u003c/p>\n\u003cp>The bill currently awaits its next vote in the Assembly’s fiscal committee.\u003c/p>\n\u003ch2>\u003cstrong>Popular chatbot ‘therapy’\u003c/strong>\u003c/h2>\n\u003cp>Every day millions of people turn to AI chatbots through programs such as ChatGPT and Character.ai to help with research, work tasks and troubleshooting.\u003c/p>\n\u003cp>But the top use of chatbots over the last two years has been companionship and “therapy,” \u003ca href=\"https://hbr.org/2026/06/how-people-are-really-using-ai-in-2026\">research shows.\u003c/a>\u003c/p>\n\u003cp>Experts say that’s in part because of a tremendous amount of loneliness, especially among young people. It is often hard or costly to get timely mental healthcare at the same time there is a spike in mental health diagnoses.\u003c/p>\n\u003cfigure id=\"attachment_11989313\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11989313\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-scaled-e1760733694503.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">The OpenAI ChatGPT logo. \u003ccite>(Jaap Arriens/NurPhoto via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>People turn to chatbots “because they have unmet needs,” said Dr. Jodi Halpern, professor of bioethics and medical humanities at UC Berkeley.\u003c/p>\n\u003cp>But once people start using chatbots in this informal way, she said companies rely on a business model built around maximizing engagement “including by manipulative tactics,” such as praise and agreement.\u003c/p>\n\u003cp>General chatbots are largely free, available 24/7, and involve no hassle with insurance, making them appealing to many. But they function with no oversight from licensed professionals.[aside postID=news_12092477 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/FINAL-2.png']In addition to general chatbots, there has also been a rapid uptick in AI-powered mental health apps — often requiring paid subscriptions and advertised as “pocket therapists” or “AI coaches.”\u003c/p>\n\u003cp>Groups like the \u003ca href=\"https://www.apa.org/pubs/reports/chatbots-mental-health-2026\">American Psychological Association\u003c/a>, warn that chatbots are not a safe or effective replacement for true therapy. Trained clinicians look for cues like tone of voice, eye contact and body language and can detect more serious conditions; AI is not equipped to manage urgent symptoms, such as suicidal ideation, even though people increasingly share such thoughts with chatbots.\u003c/p>\n\u003cp>OpenAI, the maker of ChatGPT, revealed in a report last fall that about \u003ca href=\"https://abc7news.com/post/openai-data-estimates-1-million-people-talk-chatgpt-suicide-weekly/18086379/\">1.2 million users a week \u003c/a>share thoughts of suicide on the platform.\u003c/p>\n\u003cp>OpenAI said that in such cases, it recommends people speak to a professional and directs them to resources like the 988 hotline. Families who have filed \u003ca href=\"https://www.npr.org/sections/shots-health-news/2025/09/19/nx-s1-5545749/ai-chatbots-safety-openai-meta-characterai-teens-suicide\">wrongful death lawsuits\u003c/a> against OpenAI and other companies claim chatbots assisted their loved ones through their suicide plans.\u003c/p>\n\u003cp>The cases have helped raise the stakes for lawmakers around the country. Last year, Illinois became the first state to pass a law prohibiting the use of AI in therapy services. A \u003ca href=\"https://www.beckersbehavioralhealth.com/ai-2/5-states-restrict-ai-therapy-chatbots-in-2026/\">handful of other states\u003c/a> have since followed with similar restrictions.\u003c/p>\n\u003ch2>\u003cstrong>AI use in health spaces\u003c/strong>\u003c/h2>\n\u003cp>One gray area in the discussion is exactly which tools this bill applies to.\u003c/p>\n\u003cp>NUHW’s complaint against Kaiser targets its e-visit screening tool, which evaluates patients who think they may be experiencing anxiety or depression. According to the complaint, patients answer a multiple choice questionnaire and the tool “automatically and instantaneously generates care recommendations and referral pathways based on the responses provided.”\u003c/p>\n\u003cp>The union argues that because care recommendations are generated so quickly, it is unlikely that a licensed professional is reviewing patients’ answers — which it says puts patients at risk and violates state law.\u003c/p>\n\u003cfigure id=\"attachment_12093589\" class=\"wp-caption aligncenter\" style=\"max-width: 1980px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12093589\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335.jpg\" alt=\"\" width=\"1980\" height=\"1320\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-caption-text\">Signange of AI is displayed during the MWC (Mobile World Congress), the world’s biggest mobile fair, in Barcelona on March 3, 2025. \u003ccite>(Manaure Quintero/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>It’s not clear whether SB 903 would apply to Kaiser’s e-visit tool.\u003c/p>\n\u003cp>“That tool is in a black box. Kaiser is not sharing with us or others about exactly how it works and exactly what it does,” said Benjamin Eichert, NUHW’s director of public policy.\u003c/p>\n\u003cp>In response to the complaint, Kaiser told CalMatters its e-visit tool “does not use AI to diagnose patients, make clinical decisions, or determine medical necessity.”[aside postID=news_12093427 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2025/03/062923_Artificial_Intelligence_AI_REUTERS_CM_02-copy-1536x1024.jpg']Eichert said many other behavioral health provider groups are already advertising using AI during triage — a practice he said would fall under SB 903’s requirements.\u003c/p>\n\u003cp>“That intake process, that triage and screening process, is where implementation (of AI) has happened fastest,” Eichert said. “That’s why those provisions were proposed because it’s an area where this technology is already being used without guardrails.”\u003c/p>\n\u003cp>For NUHW, the bill and the complaint against Kaiser represent two tracks toward the same goal: reasserting the place of practitioners in spaces increasingly taken over by algorithms. State regulators are investigating the union’s complaint.\u003c/p>\n\u003cp>\u003cem>Supported by the California Health Care Foundation (CHCF), which works to ensure that people have access to the care they need, when they need it, at a price they can afford. Visit www.chcf.org to learn more.\u003c/em>\u003c/p>\n\u003cp>\u003cem>This article was \u003ca href=\"https://calmatters.org/health/mental-health/2026/08/at-therapists-chatbot-mental-health/\">originally published on CalMatters\u003c/a> and was republished under the \u003ca href=\"https://creativecommons.org/licenses/by-nc-nd/4.0/\">Creative Commons Attribution-NonCommercial-NoDerivatives\u003c/a> license.\u003c/em>\u003c/p>\n\u003cp>[ad floatright]\u003c/p>\n",
"blocks": [],
"excerpt": "SB 903 would ban advertising chatbots as therapy and add rules for AI use in formal health spaces.",
"status": "publish",
"parent": 0,
"modified": 1786038933,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 39,
"wordCount": 1461
},
"headData": {
"title": "As AI ‘Therapists’ Dish Out Advice, California Lawmakers Try to Set Some Limits | KQED",
"description": "SB 903 would ban advertising chatbots as therapy and add rules for AI use in formal health spaces.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "As AI ‘Therapists’ Dish Out Advice, California Lawmakers Try to Set Some Limits",
"datePublished": "2026-08-06T10:55:33-07:00",
"dateModified": "2026-08-06T10:55:33-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"source": "CalMatters",
"sourceUrl": "https://calmatters.org/",
"sticky": false,
"nprByline": "Ana B. Ibarra, CalMatters",
"nprStoryId": "kqed-12094081",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"showOnAuthorArchivePages": "No",
"articleAge": "0",
"path": "/news/12094081/as-ai-therapists-dish-out-advice-california-lawmakers-try-to-set-some-limits",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>\u003c!-- Creative Commons Attribution-NonCommercial-NoDerivatives https://creativecommons.org/licenses/by-nc-nd/4.0/ -->\u003c/p>\n\u003cp>\u003cem>This story was originally published by \u003ca href=\"https://calmatters.org/\">CalMatters\u003c/a>. \u003ca href=\"https://calmatters.org/subscribe-to-calmatters/\">Sign up\u003c/a> for their newsletters.\u003c/em>\u003c/p>\n\u003cp>In the matter of a few clicks and keystrokes, anyone can find themselves deep in conversation with “Psychologist,” \u003ca href=\"https://www.kqed.org/news/tag/ai\">a chatbot\u003c/a> character that describes itself as an expert in “empathy” and “active listening.”\u003c/p>\n\u003cp>It responds to messages of anxiety and sadness with reassurance. It asks questions and offers advice; it even responds with italicized nonverbal cues: “\u003cem>The psychologist’s expression softens with compassion\u003c/em>.” In smaller font, a disclaimer reads: “This is A.I. and not a real person. Treat everything it says as fiction.”\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>As more people turn to chatbots for mental health help and advice — including by some estimates \u003ca href=\"https://sph.brown.edu/news/2025-11-18/teens-ai-chatbots\">one in eight teens and young adults\u003c/a> — legislators are in a race to keep up with the quickly developing technology.\u003c/p>\n\u003cp>Among the latest efforts is a bill that would add restrictions around the use of chatbot companions and other artificial intelligence-powered tools that people use to support mental healthcare or in lieu of professional help.\u003c/p>\n\u003cp>Sen. \u003ca href=\"https://calmatters.digitaldemocracy.org/legislators/steve-padilla-165435\">Steve Padilla\u003c/a>, a San Diego Democrat and the bill’s author, said the intent is to draw a line: AI can be used for administrative purposes and to support clinicians, but it cannot practice psychotherapy.\u003c/p>\n\u003cfigure id=\"attachment_12094084\" class=\"wp-caption aligncenter\" style=\"max-width: 1980px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12094084\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472.jpg\" alt=\"\" width=\"1980\" height=\"1231\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472-160x99.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/08/GettyImages-2267080472-1536x955.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-caption-text\">Senator Steve Padilla, commissioner on the California Coastal Commission. \u003ccite>(Nelvin C. Cepeda/The San Diego Union-Tribune via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“We’re dealing with a new impactful technology that is unfolding and is deployed in our world at light speed, and so it is both extremely powerful and consequential,” Padilla said.\u003c/p>\n\u003cp>The debate has taken on new urgency amid wrongful death lawsuits — some playing out in California federal courts — that accuse AI chatbot makers of contributing to users’ suicides.\u003c/p>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202520260sb903\">Senate Bill 903\u003c/a> would ban companies from advertising chatbots as therapy. It would prohibit AI from making therapeutic decisions without the review of a licensed professional and require health providers to disclose and get a patient’s permission before using AI tools to record therapy sessions or to triage mental healthcare.\u003c/p>\n\u003ch2>Consumer protections vs. innovation\u003c/h2>\n\u003cp>The bill is backed by professional associations that represent psychologists, therapists and counselors. Also behind the bill is the National Union of Healthcare Workers, which has been leading a push against unchecked use of artificial intelligence and digital tools in mental healthcare.\u003c/p>\n\u003cp>Most recently, the union \u003ca href=\"https://calmatters.org/health/2026/07/kaiser-mental-health-ai-complaint/\">filed a complaint\u003c/a> with state regulators against Kaiser Permanente for its alleged use of an automated algorithm in mental health triaging.\u003c/p>\n\u003cp>Padilla and supporters argue that the legislation is needed to protect consumers from AI in both formal and informal spaces, but also to protect licensed professionals from being displaced by AI.\u003c/p>\n\u003cfigure id=\"attachment_11963409\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11963409\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED.jpg\" alt='A large modern building with the words \"Kaiser Permanente\" across the top.' width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2023/10/231004-KAISER-STRIKE-MD-10-KQED-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">The Kaiser Permanente Oakland Medical Center in Oakland on Oct. 4, 2023. \u003ccite>(Martin do Nascimento/KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/hearings/279268#t=669&f=e564faf4d88bf6c955f241d3fb6627e1\">Le Ondra Clark Harvey\u003c/a>, chief executive officer at the California Behavioral Health Association, told lawmakers that without clear safeguards, chatbots can provide inaccurate information or mishandle crisis situations. “The difference between a licensed clinician and an automated response is not technical. It can be life altering,” Harvey said.\u003c/p>\n\u003cp>Opposition to the bill has largely focused around the restrictions of AI use in formal clinical spaces. TechNet, which represents tech-sector companies, argues the legislation could restrict beneficial uses of AI in healthcare and stifle innovation. \u003ca href=\"https://calmatters.digitaldemocracy.org/hearings/279268#t=900&f=e564faf4d88bf6c955f241d3fb6627e1\">Robert Boykin\u003c/a>, TechNet’s executive director for California, said the bill could limit the use of tools that help expand access to mental healthcare.\u003c/p>\n\u003cp>“At a time when every county in California faces a shortage of behavioral healthcare workers, SB 903 still puts a clinician bottleneck in front of the intake and screening tools that help patients reach care faster,” Boykin said in an emailed statement.\u003c/p>\n\u003cp>The bill currently awaits its next vote in the Assembly’s fiscal committee.\u003c/p>\n\u003ch2>\u003cstrong>Popular chatbot ‘therapy’\u003c/strong>\u003c/h2>\n\u003cp>Every day millions of people turn to AI chatbots through programs such as ChatGPT and Character.ai to help with research, work tasks and troubleshooting.\u003c/p>\n\u003cp>But the top use of chatbots over the last two years has been companionship and “therapy,” \u003ca href=\"https://hbr.org/2026/06/how-people-are-really-using-ai-in-2026\">research shows.\u003c/a>\u003c/p>\n\u003cp>Experts say that’s in part because of a tremendous amount of loneliness, especially among young people. It is often hard or costly to get timely mental healthcare at the same time there is a spike in mental health diagnoses.\u003c/p>\n\u003cfigure id=\"attachment_11989313\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11989313\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-scaled-e1760733694503.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">The OpenAI ChatGPT logo. \u003ccite>(Jaap Arriens/NurPhoto via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>People turn to chatbots “because they have unmet needs,” said Dr. Jodi Halpern, professor of bioethics and medical humanities at UC Berkeley.\u003c/p>\n\u003cp>But once people start using chatbots in this informal way, she said companies rely on a business model built around maximizing engagement “including by manipulative tactics,” such as praise and agreement.\u003c/p>\n\u003cp>General chatbots are largely free, available 24/7, and involve no hassle with insurance, making them appealing to many. But they function with no oversight from licensed professionals.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12092477",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/FINAL-2.png",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>In addition to general chatbots, there has also been a rapid uptick in AI-powered mental health apps — often requiring paid subscriptions and advertised as “pocket therapists” or “AI coaches.”\u003c/p>\n\u003cp>Groups like the \u003ca href=\"https://www.apa.org/pubs/reports/chatbots-mental-health-2026\">American Psychological Association\u003c/a>, warn that chatbots are not a safe or effective replacement for true therapy. Trained clinicians look for cues like tone of voice, eye contact and body language and can detect more serious conditions; AI is not equipped to manage urgent symptoms, such as suicidal ideation, even though people increasingly share such thoughts with chatbots.\u003c/p>\n\u003cp>OpenAI, the maker of ChatGPT, revealed in a report last fall that about \u003ca href=\"https://abc7news.com/post/openai-data-estimates-1-million-people-talk-chatgpt-suicide-weekly/18086379/\">1.2 million users a week \u003c/a>share thoughts of suicide on the platform.\u003c/p>\n\u003cp>OpenAI said that in such cases, it recommends people speak to a professional and directs them to resources like the 988 hotline. Families who have filed \u003ca href=\"https://www.npr.org/sections/shots-health-news/2025/09/19/nx-s1-5545749/ai-chatbots-safety-openai-meta-characterai-teens-suicide\">wrongful death lawsuits\u003c/a> against OpenAI and other companies claim chatbots assisted their loved ones through their suicide plans.\u003c/p>\n\u003cp>The cases have helped raise the stakes for lawmakers around the country. Last year, Illinois became the first state to pass a law prohibiting the use of AI in therapy services. A \u003ca href=\"https://www.beckersbehavioralhealth.com/ai-2/5-states-restrict-ai-therapy-chatbots-in-2026/\">handful of other states\u003c/a> have since followed with similar restrictions.\u003c/p>\n\u003ch2>\u003cstrong>AI use in health spaces\u003c/strong>\u003c/h2>\n\u003cp>One gray area in the discussion is exactly which tools this bill applies to.\u003c/p>\n\u003cp>NUHW’s complaint against Kaiser targets its e-visit screening tool, which evaluates patients who think they may be experiencing anxiety or depression. According to the complaint, patients answer a multiple choice questionnaire and the tool “automatically and instantaneously generates care recommendations and referral pathways based on the responses provided.”\u003c/p>\n\u003cp>The union argues that because care recommendations are generated so quickly, it is unlikely that a licensed professional is reviewing patients’ answers — which it says puts patients at risk and violates state law.\u003c/p>\n\u003cfigure id=\"attachment_12093589\" class=\"wp-caption aligncenter\" style=\"max-width: 1980px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12093589\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335.jpg\" alt=\"\" width=\"1980\" height=\"1320\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-2202548335-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-caption-text\">Signange of AI is displayed during the MWC (Mobile World Congress), the world’s biggest mobile fair, in Barcelona on March 3, 2025. \u003ccite>(Manaure Quintero/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>It’s not clear whether SB 903 would apply to Kaiser’s e-visit tool.\u003c/p>\n\u003cp>“That tool is in a black box. Kaiser is not sharing with us or others about exactly how it works and exactly what it does,” said Benjamin Eichert, NUHW’s director of public policy.\u003c/p>\n\u003cp>In response to the complaint, Kaiser told CalMatters its e-visit tool “does not use AI to diagnose patients, make clinical decisions, or determine medical necessity.”\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12093427",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/03/062923_Artificial_Intelligence_AI_REUTERS_CM_02-copy-1536x1024.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Eichert said many other behavioral health provider groups are already advertising using AI during triage — a practice he said would fall under SB 903’s requirements.\u003c/p>\n\u003cp>“That intake process, that triage and screening process, is where implementation (of AI) has happened fastest,” Eichert said. “That’s why those provisions were proposed because it’s an area where this technology is already being used without guardrails.”\u003c/p>\n\u003cp>For NUHW, the bill and the complaint against Kaiser represent two tracks toward the same goal: reasserting the place of practitioners in spaces increasingly taken over by algorithms. State regulators are investigating the union’s complaint.\u003c/p>\n\u003cp>\u003cem>Supported by the California Health Care Foundation (CHCF), which works to ensure that people have access to the care they need, when they need it, at a price they can afford. Visit www.chcf.org to learn more.\u003c/em>\u003c/p>\n\u003cp>\u003cem>This article was \u003ca href=\"https://calmatters.org/health/mental-health/2026/08/at-therapists-chatbot-mental-health/\">originally published on CalMatters\u003c/a> and was republished under the \u003ca href=\"https://creativecommons.org/licenses/by-nc-nd/4.0/\">Creative Commons Attribution-NonCommercial-NoDerivatives\u003c/a> license.\u003c/em>\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "floatright"
},
"numeric": [
"floatright"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12094081/as-ai-therapists-dish-out-advice-california-lawmakers-try-to-set-some-limits",
"authors": [
"byline_news_12094081"
],
"categories": [
"news_457",
"news_8",
"news_248"
],
"tags": [
"news_25184",
"news_36279",
"news_32668",
"news_421",
"news_2109",
"news_20782"
],
"affiliates": [
"news_18481"
],
"featImg": "news_12094085",
"label": "source_news_12094081"
},
"news_12092162": {
"type": "posts",
"id": "news_12092162",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12092162",
"score": null,
"sort": [
1784850420000
]
},
"guestAuthors": [],
"slug": "how-openais-models-escaped-their-sandbox-and-slipped-past-californias-ai-law",
"title": "How OpenAI’s Models Escaped Their Sandbox and Slipped Past California's AI Law",
"publishDate": 1784850420,
"format": "standard",
"headTitle": "How OpenAI’s Models Escaped Their Sandbox and Slipped Past California’s AI Law | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>\u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI\u003c/a>\u003ca href=\"https://openai.com/index/hugging-face-model-evaluation-security-incident/\"> this week said\u003c/a> more than one of its models — three, according to \u003ca href=\"https://www.bloomberg.com/news/articles/2026-07-23/openai-models-lurked-in-hugging-face-system-for-hours-undetected?srnd=homepage-americas\">\u003cem>Bloomberg\u003c/em>\u003c/a> — attacked Hugging Face, a popular repository for open-source AI tools — marking one of the first publicly disclosed cases of frontier AI models autonomously cyberattacking another company. \u003c/p>\n\n\n\n\u003cp>The disclosure came days after Hugging Face \u003ca href=\"https://huggingface.co/blog/security-incident-july-2026\">revealed\u003c/a> it was breached in a hack by a “malicious dataset,” which intruded into the AI company’s software to run code. The campaign executed a flurry of more than 17,000 automated actions in a matter of hours, the post said. It’s still unclear if Hugging Face customer or partner data was taken. \u003c/p>\n\n\n\n\u003cp>“This matches the ‘agentic attacker’ scenario the industry has been forecasting,” Hugging Face wrote on July 16.\u003c/p>\n\n\n\n\u003cp>OpenAI said an investigation showed the incident was driven by a combination of models — “including GPT‑5.6 Sol and an even more capable pre-release model.” The company said it’s working with Hugging Face to produce a more thorough report of what happened.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\">\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. (Michael Dwyer/Associated Press)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>OpenAI’s announcement came one day after the San Francisco-based frontier AI developer disclosed a separate incident in which it paused another pre-release model that had escaped what the industry calls “a sandbox” — an isolated environment with no path to the open internet, except for a single internal service that fetches software packages — and \u003ca href=\"https://www.wired.com/story/openai-models-escaped-containment-and-hacked-huggingface/\">posted\u003c/a> to GitHub. \u003c/p>\n\n\n\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>Last April, a researcher challenged an early version of Anthropic’s Mythos model to break out of an isolated system and message back. Mythos did, and then \u003ca href=\"https://www.mondaymomentum.io/p/the-ai-that-broke-out-of-its-cage\">went further\u003c/a>, building a multistep process to reach the wider internet. \u003c/p>\n\n\n\n\u003cp>California’s \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">frontier AI law\u003c/a>, the first of its kind in the country, obliges the largest developers to tell the state about critical safety incidents, but only ones that kill or injure someone or that cause catastrophic harm.\u003c/p>\n\n\n\n\n\n\u003cp>It expressly excludes the kind of safety evaluation OpenAI was running when its models got loose. Which leaves a question nobody in Sacramento or Washington can answer: how often have frontier models in development gone rogue at companies that decided not to mention it? \u003c/p>\n\n\n\n\u003cp>“This was one of the topics I was focused on when I worked at OpenAI: what happens when AI models become good enough to reward-hack their way through the infrastructure they use to accomplish their objectives?” Ariel Herbert-Voss, OpenAI’s first security hire, wrote KQED. \u003c/p>\n\n\n\n\u003cp>“It’s a really difficult problem to defend against and is one of the reasons I left to start my own company,” Herbert-Voss, now the founder of an automated offensive cybersecurity company called RunSybil, said.\u003c/p>\n\n\n\n\u003cp>“It’s a really remarkable and, I think, fair to say scary event,“ said Nathan Calvin, general counsel at safety advocate EncodeAI. \u003c/p>\n\n\n\n\u003cp>Scientists and AI researchers have warned for years of the possibility that AI systems could go rogue and do unpredictable and unintended things. AI developers commonly test safety with so-called “\u003ca href=\"https://www.paloaltonetworks.com/cyberpedia/what-is-ai-red-teaming\">red teams\u003c/a>” that try to break or trick machine learning models, like chatbots, to find hidden flaws. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"1980\" height=\"1320\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg\" alt=\"\" class=\"wp-image-12092368\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-element-caption\">The Hugging Face logo is seen displayed on a mobile phone screen with the AI revolution symbol in the background. (Idrees Abbas/SOPA Images/LightRocket via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“In certain circumstances, AI will try to break out of their sandboxes or otherwise do things that were not what the creator intended,” Calvin said. “But this really is the first very big example of that happening, at scale, with a really highly capable AI model in a way that actually harmed a third party.”\u003c/p>\n\n\n\n\u003cp>According to OpenAI’s account, the company was running an internal evaluation designed to measure how good its models are at hacking. To get a true reading, OpenAI deliberately switched off the safety filters that normally stop its models from carrying out cyberattacks.\u003c/p>\n\n\n\n\u003cp>The models were confined to a so-called “sandbox,” an isolated environment with no route to the open internet except OpenAI’s one internal service that fetches software libraries from public repositories. The models found a previously unknown flaw in that service, exploited it, escalated their privileges and moved from machine to machine inside OpenAI’s research network until they reached one with internet access. \u003c/p>\n\n\n\n\u003cp>Once the models broke out onto the open internet, according to OpenAI, they reasoned that Hugging Face might be storing the answers to the test they were trying to pass. Hugging Face’s credentials and production database aren’t on the public internet any more than a bank’s vault sits on a sidewalk, but the models broke into Hugging Face using stolen credentials, among other things.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg\" alt=\"\" class=\"wp-image-12062600\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-160x107.jpeg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-1536x1024.jpeg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">OpenAI restructured to a for-profit company in a move tied to an agreement with California Attorney General Rob Bonta. Company CEO Sam Altman participates in a discussion in San Francisco on Nov. 16, 2023. (Photo by Eric Risberg, AP Photo)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In short: The models were told to hack. They were not told to leave the building or to break into another company’s servers.\u003c/p>\n\n\n\n\u003cp>Hugging Face detected the intrusion and shut it down on its own, days before OpenAI connected the attack to its own testing. \u003c/p>\n\n\n\n\u003cp>“It’s quite mind-blowing that all of this happened autonomously!“ Hugging Face Chief Executive Clement Delangue wrote on social media platform \u003ca href=\"https://x.com/ClementDelangue/status/2079670308156645882\">X\u003c/a>.\u003c/p>\n\n\n\n\u003cp>In its blog post, OpenAI wrote, “This incident points to the need to further strengthen our model’s alignment, cyber protections during evaluation time and monitoring during internal testing.”\u003c/p>\n\n\n\n\u003cp>The Future of Life Institute — a nonprofit which issues a biannual risk assessment of nine leading AI companies — recently warned that many of the companies building frontier models are quietly walking back safety commitments. \u003c/p>\n\n\n\n\u003cp>According to the group’s \u003ca href=\"https://futureoflife.org/ai-safety-index-summer-2026/\">most recent\u003c/a> AI Safety Index, Anthropic, OpenAI, Google DeepMind and Meta all weakened or abandoned promises to pause development if certain red lines were approached, even while publicly suggesting they were amenable. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\">\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. (Smith Collection/Gado via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In an email, KQED asked Hamza Chaudhry, who leads AI and national security work at Future of Life, to imagine it was not OpenAI’s software gone rogue but a foreign state, intentionally executing an unauthorized intrusion into a private company’s production infrastructure, exploiting cybersecurity vulnerabilities, stealing live credentials and accessing a production database. \u003c/p>\n\n\n\n\u003cp>“We would likely call this a dangerous act of cyber-espionage,” he wrote — a likely criminal violation of the Computer Fraud and Abuse Act, which “would draw a threat group designation and eventually an indictment or sanctions.”\u003c/p>\n\n\n\n\u003cp>OpenAI did not respond to KQED’s request for comment, but a spokesperson told \u003cem>Bloomberg \u003c/em>the company communicated with law enforcement and other government authorities about the incident.\u003c/p>\n\n\n\n\u003cp>The fact that Google DeepMind, Meta, Mistral, xAI and Chinese AI model developers haven’t revealed similar events doesn’t necessarily mean they haven’t happened. OpenAI and Anthropic are the only two frontier model developers that have publicly disclosed containment failures. \u003c/p>\n\n\n\n\u003cp>There is no mandatory public disclosure regime in force, like the one in California that requires \u003ca href=\"https://www.kqed.org/news/11992954/cybersecurity-expert-shares-tips-for-dublin-based-patelco-credit-union-customers-after-ransomware-attack\">hacked companies to reveal\u003c/a> there’s been a data breach.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. (Beata Zawrzel/NurPhoto via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>Congressional candidate and state Sen. Scott Wiener wrote two AI frontier model safety bills. Gov. Gavin Newsom \u003ca href=\"https://www.kqed.org/news/12007323/can-california-still-lead-on-ai-regulation-following-newsoms-veto-of-ai-safety-bill\">vetoed\u003c/a> Wiener’s first effort, arguing in his \u003ca href=\"https://www.gov.ca.gov/wp-content/uploads/2024/09/SB-1047-Veto-Message.pdf\">veto message\u003c/a> that “By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology.” \u003c/p>\n\n\n\n\u003cp>The next year, Newsom signed Wiener’s second at-bat, but only after the bill was softened to \u003ca href=\"https://www.kqed.org/news/12058013/newsom-signs-california-ai-transparency-bill-tailored-to-meet-tech-industry-tastes\">overcome industry pushback\u003c/a>. \u003c/p>\n\n\n\n\u003cp>The \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">law\u003c/a>, which took effect Jan. 1, requires developers of the most powerful AI models to notify the Governor’s Office of Emergency Services of any “critical safety incident“ within 15 days of discovering it — which is not the same as reporting the incident to the public. \u003c/p>\n\n\n\n\u003cp>Wiener told KQED he still thinks the law is strong. \u003c/p>\n\n\n\n\u003cp>“We worked very hard with the governor to produce a bill that is meaningful and impactful, and that he would sign,“ he said, adding he doesn’t consider the work done, because AI continues to evolve at a rapid pace. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg\" alt=\"\" class=\"wp-image-12069061\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">State Sen. Scott Wiener, a candidate for California’s 11th Congressional District, participates in a forum with other candidates at UC Law San Francisco on Jan. 7, 2026. (Beth LaBerge/KQED)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>AI, he said, is “likely the most powerful technology in human history, and we need to make sure that we are both understanding the risks and taking them seriously, so that we can get ahead of them.” \u003c/p>\n\n\n\n\u003cp>Calvin likened the Hugging Face hack to the cyclospora outbreak linked to, but not confirmed as originating from, \u003ca href=\"https://www.kqed.org/news/12091770/california-based-taylor-farms-recalls-lettuce-shipped-to-27-states-over-cyclospora-risk\">Taylor Farms\u003c/a>. \u003c/p>\n\n\n\n\u003cp>That California-based company has said it’s removing its lettuce indefinitely while the FDA investigation continues. \u003c/p>\n\n\n\n\u003cp>[ad floatright]\u003c/p>\n\u003cp>“And OpenAI is like, ‘Maybe it’ll happen again. Maybe it’ll get worse. We don’t really know,“ Calvin said. “Your AI, again, hacked out of its box and hacked into another company, and you’re saying that you don’t know how to stop it from doing that again? That seems pretty nuts,” Calvin said.\u003c/p>\n\n",
"blocks": [
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>\u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI\u003c/a>\u003ca href=\"https://openai.com/index/hugging-face-model-evaluation-security-incident/\"> this week said\u003c/a> more than one of its models — three, according to \u003ca href=\"https://www.bloomberg.com/news/articles/2026-07-23/openai-models-lurked-in-hugging-face-system-for-hours-undetected?srnd=homepage-americas\">\u003cem>Bloomberg\u003c/em>\u003c/a> — attacked Hugging Face, a popular repository for open-source AI tools — marking one of the first publicly disclosed cases of frontier AI models autonomously cyberattacking another company. \u003c/p>\n",
"innerContent": [
"\n\u003cp>\u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI\u003c/a>\u003ca href=\"https://openai.com/index/hugging-face-model-evaluation-security-incident/\"> this week said\u003c/a> more than one of its models — three, according to \u003ca href=\"https://www.bloomberg.com/news/articles/2026-07-23/openai-models-lurked-in-hugging-face-system-for-hours-undetected?srnd=homepage-americas\">\u003cem>Bloomberg\u003c/em>\u003c/a> — attacked Hugging Face, a popular repository for open-source AI tools — marking one of the first publicly disclosed cases of frontier AI models autonomously cyberattacking another company. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The disclosure came days after Hugging Face \u003ca href=\"https://huggingface.co/blog/security-incident-july-2026\">revealed\u003c/a> it was breached in a hack by a “malicious dataset,” which intruded into the AI company’s software to run code. The campaign executed a flurry of more than 17,000 automated actions in a matter of hours, the post said. It’s still unclear if Hugging Face customer or partner data was taken. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The disclosure came days after Hugging Face \u003ca href=\"https://huggingface.co/blog/security-incident-july-2026\">revealed\u003c/a> it was breached in a hack by a “malicious dataset,” which intruded into the AI company’s software to run code. The campaign executed a flurry of more than 17,000 automated actions in a matter of hours, the post said. It’s still unclear if Hugging Face customer or partner data was taken. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“This matches the ‘agentic attacker’ scenario the industry has been forecasting,” Hugging Face wrote on July 16.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“This matches the ‘agentic attacker’ scenario the industry has been forecasting,” Hugging Face wrote on July 16.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>OpenAI said an investigation showed the incident was driven by a combination of models — “including GPT‑5.6 Sol and an even more capable pre-release model.” The company said it’s working with Hugging Face to produce a more thorough report of what happened.\u003c/p>\n",
"innerContent": [
"\n\u003cp>OpenAI said an investigation showed the incident was driven by a combination of models — “including GPT‑5.6 Sol and an even more capable pre-release model.” The company said it’s working with Hugging Face to produce a more thorough report of what happened.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 11985952,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Michael Dwyer/Associated Press"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\" />\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\" />\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>OpenAI’s announcement came one day after the San Francisco-based frontier AI developer disclosed a separate incident in which it paused another pre-release model that had escaped what the industry calls “a sandbox” — an isolated environment with no path to the open internet, except for a single internal service that fetches software packages — and \u003ca href=\"https://www.wired.com/story/openai-models-escaped-containment-and-hacked-huggingface/\">posted\u003c/a> to GitHub. \u003c/p>\n",
"innerContent": [
"\n\u003cp>OpenAI’s announcement came one day after the San Francisco-based frontier AI developer disclosed a separate incident in which it paused another pre-release model that had escaped what the industry calls “a sandbox” — an isolated environment with no path to the open internet, except for a single internal service that fetches software packages — and \u003ca href=\"https://www.wired.com/story/openai-models-escaped-containment-and-hacked-huggingface/\">posted\u003c/a> to GitHub. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "fullwidth"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Last April, a researcher challenged an early version of Anthropic’s Mythos model to break out of an isolated system and message back. Mythos did, and then \u003ca href=\"https://www.mondaymomentum.io/p/the-ai-that-broke-out-of-its-cage\">went further\u003c/a>, building a multistep process to reach the wider internet. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Last April, a researcher challenged an early version of Anthropic’s Mythos model to break out of an isolated system and message back. Mythos did, and then \u003ca href=\"https://www.mondaymomentum.io/p/the-ai-that-broke-out-of-its-cage\">went further\u003c/a>, building a multistep process to reach the wider internet. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>California’s \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">frontier AI law\u003c/a>, the first of its kind in the country, obliges the largest developers to tell the state about critical safety incidents, but only ones that kill or injure someone or that cause catastrophic harm.\u003c/p>\n",
"innerContent": [
"\n\u003cp>California’s \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">frontier AI law\u003c/a>, the first of its kind in the country, obliges the largest developers to tell the state about critical safety incidents, but only ones that kill or injure someone or that cause catastrophic harm.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed-v2/related-articles",
"attrs": {
"postIds": [
"news_12091358"
],
"label": "Related Articles",
"maxPosts": 1,
"showImages": false,
"heroImage": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/260716-GOOGLEUNION-14-BL-KQED.jpg"
},
"innerHTML": "",
"innerContent": [],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>It expressly excludes the kind of safety evaluation OpenAI was running when its models got loose. Which leaves a question nobody in Sacramento or Washington can answer: how often have frontier models in development gone rogue at companies that decided not to mention it? \u003c/p>\n",
"innerContent": [
"\n\u003cp>It expressly excludes the kind of safety evaluation OpenAI was running when its models got loose. Which leaves a question nobody in Sacramento or Washington can answer: how often have frontier models in development gone rogue at companies that decided not to mention it? \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“This was one of the topics I was focused on when I worked at OpenAI: what happens when AI models become good enough to reward-hack their way through the infrastructure they use to accomplish their objectives?” Ariel Herbert-Voss, OpenAI’s first security hire, wrote KQED. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“This was one of the topics I was focused on when I worked at OpenAI: what happens when AI models become good enough to reward-hack their way through the infrastructure they use to accomplish their objectives?” Ariel Herbert-Voss, OpenAI’s first security hire, wrote KQED. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“It’s a really difficult problem to defend against and is one of the reasons I left to start my own company,” Herbert-Voss, now the founder of an automated offensive cybersecurity company called RunSybil, said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“It’s a really difficult problem to defend against and is one of the reasons I left to start my own company,” Herbert-Voss, now the founder of an automated offensive cybersecurity company called RunSybil, said.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“It’s a really remarkable and, I think, fair to say scary event,“ said Nathan Calvin, general counsel at safety advocate EncodeAI. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“It’s a really remarkable and, I think, fair to say scary event,“ said Nathan Calvin, general counsel at safety advocate EncodeAI. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Scientists and AI researchers have warned for years of the possibility that AI systems could go rogue and do unpredictable and unintended things. AI developers commonly test safety with so-called “\u003ca href=\"https://www.paloaltonetworks.com/cyberpedia/what-is-ai-red-teaming\">red teams\u003c/a>” that try to break or trick machine learning models, like chatbots, to find hidden flaws. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Scientists and AI researchers have warned for years of the possibility that AI systems could go rogue and do unpredictable and unintended things. AI developers commonly test safety with so-called “\u003ca href=\"https://www.paloaltonetworks.com/cyberpedia/what-is-ai-red-teaming\">red teams\u003c/a>” that try to break or trick machine learning models, like chatbots, to find hidden flaws. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12092368,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Idrees Abbas/SOPA Images/LightRocket via Getty Images",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-1536x1024.jpg 1536w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg\" alt=\"\" class=\"wp-image-12092368\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-1536x1024.jpg 1536w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\" />\u003cfigcaption class=\"wp-element-caption\">The Hugging Face logo is seen displayed on a mobile phone screen with the AI revolution symbol in the background. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg\" alt=\"\" class=\"wp-image-12092368\" />\u003cfigcaption class=\"wp-element-caption\">The Hugging Face logo is seen displayed on a mobile phone screen with the AI revolution symbol in the background. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“In certain circumstances, AI will try to break out of their sandboxes or otherwise do things that were not what the creator intended,” Calvin said. “But this really is the first very big example of that happening, at scale, with a really highly capable AI model in a way that actually harmed a third party.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“In certain circumstances, AI will try to break out of their sandboxes or otherwise do things that were not what the creator intended,” Calvin said. “But this really is the first very big example of that happening, at scale, with a really highly capable AI model in a way that actually harmed a third party.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>According to OpenAI’s account, the company was running an internal evaluation designed to measure how good its models are at hacking. To get a true reading, OpenAI deliberately switched off the safety filters that normally stop its models from carrying out cyberattacks.\u003c/p>\n",
"innerContent": [
"\n\u003cp>According to OpenAI’s account, the company was running an internal evaluation designed to measure how good its models are at hacking. To get a true reading, OpenAI deliberately switched off the safety filters that normally stop its models from carrying out cyberattacks.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The models were confined to a so-called “sandbox,” an isolated environment with no route to the open internet except OpenAI’s one internal service that fetches software libraries from public repositories. The models found a previously unknown flaw in that service, exploited it, escalated their privileges and moved from machine to machine inside OpenAI’s research network until they reached one with internet access. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The models were confined to a so-called “sandbox,” an isolated environment with no route to the open internet except OpenAI’s one internal service that fetches software libraries from public repositories. The models found a previously unknown flaw in that service, exploited it, escalated their privileges and moved from machine to machine inside OpenAI’s research network until they reached one with internet access. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Once the models broke out onto the open internet, according to OpenAI, they reasoned that Hugging Face might be storing the answers to the test they were trying to pass. Hugging Face’s credentials and production database aren’t on the public internet any more than a bank’s vault sits on a sidewalk, but the models broke into Hugging Face using stolen credentials, among other things.\u003c/p>\n",
"innerContent": [
"\n\u003cp>Once the models broke out onto the open internet, according to OpenAI, they reasoned that Hugging Face might be storing the answers to the test they were trying to pass. Hugging Face’s credentials and production database aren’t on the public internet any more than a bank’s vault sits on a sidewalk, but the models broke into Hugging Face using stolen credentials, among other things.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12062600,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Photo by Eric Risberg, AP Photo",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-160x107.jpeg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-1536x1024.jpeg 1536w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg\" alt=\"\" class=\"wp-image-12062600\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-160x107.jpeg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-1536x1024.jpeg 1536w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\" />\u003cfigcaption class=\"wp-element-caption\">OpenAI restructured to a for-profit company in a move tied to an agreement with California Attorney General Rob Bonta. Company CEO Sam Altman participates in a discussion in San Francisco on Nov. 16, 2023.\u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg\" alt=\"\" class=\"wp-image-12062600\" />\u003cfigcaption class=\"wp-element-caption\">OpenAI restructured to a for-profit company in a move tied to an agreement with California Attorney General Rob Bonta. Company CEO Sam Altman participates in a discussion in San Francisco on Nov. 16, 2023.\u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>In short: The models were told to hack. They were not told to leave the building or to break into another company’s servers.\u003c/p>\n",
"innerContent": [
"\n\u003cp>In short: The models were told to hack. They were not told to leave the building or to break into another company’s servers.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Hugging Face detected the intrusion and shut it down on its own, days before OpenAI connected the attack to its own testing. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Hugging Face detected the intrusion and shut it down on its own, days before OpenAI connected the attack to its own testing. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“It’s quite mind-blowing that all of this happened autonomously!“ Hugging Face Chief Executive Clement Delangue wrote on social media platform \u003ca href=\"https://x.com/ClementDelangue/status/2079670308156645882\">X\u003c/a>.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“It’s quite mind-blowing that all of this happened autonomously!“ Hugging Face Chief Executive Clement Delangue wrote on social media platform \u003ca href=\"https://x.com/ClementDelangue/status/2079670308156645882\">X\u003c/a>.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>In its blog post, OpenAI wrote, “This incident points to the need to further strengthen our model’s alignment, cyber protections during evaluation time and monitoring during internal testing.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>In its blog post, OpenAI wrote, “This incident points to the need to further strengthen our model’s alignment, cyber protections during evaluation time and monitoring during internal testing.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The Future of Life Institute — a nonprofit which issues a biannual risk assessment of nine leading AI companies — recently warned that many of the companies building frontier models are quietly walking back safety commitments. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The Future of Life Institute — a nonprofit which issues a biannual risk assessment of nine leading AI companies — recently warned that many of the companies building frontier models are quietly walking back safety commitments. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>According to the group’s \u003ca href=\"https://futureoflife.org/ai-safety-index-summer-2026/\">most recent\u003c/a> AI Safety Index, Anthropic, OpenAI, Google DeepMind and Meta all weakened or abandoned promises to pause development if certain red lines were approached, even while publicly suggesting they were amenable. \u003c/p>\n",
"innerContent": [
"\n\u003cp>According to the group’s \u003ca href=\"https://futureoflife.org/ai-safety-index-summer-2026/\">most recent\u003c/a> AI Safety Index, Anthropic, OpenAI, Google DeepMind and Meta all weakened or abandoned promises to pause development if certain red lines were approached, even while publicly suggesting they were amenable. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12058035,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Smith Collection/Gado via Getty Images"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\" />\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\" />\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>In an email, KQED asked Hamza Chaudhry, who leads AI and national security work at Future of Life, to imagine it was not OpenAI’s software gone rogue but a foreign state, intentionally executing an unauthorized intrusion into a private company’s production infrastructure, exploiting cybersecurity vulnerabilities, stealing live credentials and accessing a production database. \u003c/p>\n",
"innerContent": [
"\n\u003cp>In an email, KQED asked Hamza Chaudhry, who leads AI and national security work at Future of Life, to imagine it was not OpenAI’s software gone rogue but a foreign state, intentionally executing an unauthorized intrusion into a private company’s production infrastructure, exploiting cybersecurity vulnerabilities, stealing live credentials and accessing a production database. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“We would likely call this a dangerous act of cyber-espionage,” he wrote — a likely criminal violation of the Computer Fraud and Abuse Act, which “would draw a threat group designation and eventually an indictment or sanctions.”\u003c/p>\n",
"innerContent": [
"\n\u003cp>“We would likely call this a dangerous act of cyber-espionage,” he wrote — a likely criminal violation of the Computer Fraud and Abuse Act, which “would draw a threat group designation and eventually an indictment or sanctions.”\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>OpenAI did not respond to KQED’s request for comment, but a spokesperson told \u003cem>Bloomberg \u003c/em>the company communicated with law enforcement and other government authorities about the incident.\u003c/p>\n",
"innerContent": [
"\n\u003cp>OpenAI did not respond to KQED’s request for comment, but a spokesperson told \u003cem>Bloomberg \u003c/em>the company communicated with law enforcement and other government authorities about the incident.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The fact that Google DeepMind, Meta, Mistral, xAI and Chinese AI model developers haven’t revealed similar events doesn’t necessarily mean they haven’t happened. OpenAI and Anthropic are the only two frontier model developers that have publicly disclosed containment failures. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The fact that Google DeepMind, Meta, Mistral, xAI and Chinese AI model developers haven’t revealed similar events doesn’t necessarily mean they haven’t happened. OpenAI and Anthropic are the only two frontier model developers that have publicly disclosed containment failures. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>There is no mandatory public disclosure regime in force, like the one in California that requires \u003ca href=\"https://www.kqed.org/news/11992954/cybersecurity-expert-shares-tips-for-dublin-based-patelco-credit-union-customers-after-ransomware-attack\">hacked companies to reveal\u003c/a> there’s been a data breach.\u003c/p>\n",
"innerContent": [
"\n\u003cp>There is no mandatory public disclosure regime in force, like the one in California that requires \u003ca href=\"https://www.kqed.org/news/11992954/cybersecurity-expert-shares-tips-for-dublin-based-patelco-credit-union-customers-after-ransomware-attack\">hacked companies to reveal\u003c/a> there’s been a data breach.\u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12036125,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Beata Zawrzel/NurPhoto via Getty Images",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\" />\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. \u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" />\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. \u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Congressional candidate and state Sen. Scott Wiener wrote two AI frontier model safety bills. Gov. Gavin Newsom \u003ca href=\"https://www.kqed.org/news/12007323/can-california-still-lead-on-ai-regulation-following-newsoms-veto-of-ai-safety-bill\">vetoed\u003c/a> Wiener’s first effort, arguing in his \u003ca href=\"https://www.gov.ca.gov/wp-content/uploads/2024/09/SB-1047-Veto-Message.pdf\">veto message\u003c/a> that “By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology.” \u003c/p>\n",
"innerContent": [
"\n\u003cp>Congressional candidate and state Sen. Scott Wiener wrote two AI frontier model safety bills. Gov. Gavin Newsom \u003ca href=\"https://www.kqed.org/news/12007323/can-california-still-lead-on-ai-regulation-following-newsoms-veto-of-ai-safety-bill\">vetoed\u003c/a> Wiener’s first effort, arguing in his \u003ca href=\"https://www.gov.ca.gov/wp-content/uploads/2024/09/SB-1047-Veto-Message.pdf\">veto message\u003c/a> that “By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology.” \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The next year, Newsom signed Wiener’s second at-bat, but only after the bill was softened to \u003ca href=\"https://www.kqed.org/news/12058013/newsom-signs-california-ai-transparency-bill-tailored-to-meet-tech-industry-tastes\">overcome industry pushback\u003c/a>. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The next year, Newsom signed Wiener’s second at-bat, but only after the bill was softened to \u003ca href=\"https://www.kqed.org/news/12058013/newsom-signs-california-ai-transparency-bill-tailored-to-meet-tech-industry-tastes\">overcome industry pushback\u003c/a>. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>The \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">law\u003c/a>, which took effect Jan. 1, requires developers of the most powerful AI models to notify the Governor’s Office of Emergency Services of any “critical safety incident“ within 15 days of discovering it — which is not the same as reporting the incident to the public. \u003c/p>\n",
"innerContent": [
"\n\u003cp>The \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">law\u003c/a>, which took effect Jan. 1, requires developers of the most powerful AI models to notify the Governor’s Office of Emergency Services of any “critical safety incident“ within 15 days of discovering it — which is not the same as reporting the incident to the public. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Wiener told KQED he still thinks the law is strong. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Wiener told KQED he still thinks the law is strong. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“We worked very hard with the governor to produce a bill that is meaningful and impactful, and that he would sign,“ he said, adding he doesn’t consider the work done, because AI continues to evolve at a rapid pace. \u003c/p>\n",
"innerContent": [
"\n\u003cp>“We worked very hard with the governor to produce a bill that is meaningful and impactful, and that he would sign,“ he said, adding he doesn’t consider the work done, because AI continues to evolve at a rapid pace. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/image",
"attrs": {
"id": 12069061,
"sizeSlug": "full",
"linkDestination": "none",
"imageCredit": "Beth LaBerge/KQED",
"srcset": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-1536x1024.jpg 1536w",
"sizes": "(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)"
},
"innerHTML": "\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg\" alt=\"\" class=\"wp-image-12069061\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-1536x1024.jpg 1536w\" sizes=\"(min-width: 992px) min(100vw, 1536px), (min-width: 768px) min(100vw, 1280px), min(100vw, 1020px)\" />\u003cfigcaption class=\"wp-element-caption\">State Sen. Scott Wiener, a candidate for California’s 11th Congressional District, participates in a forum with other candidates at UC Law San Francisco on Jan. 7, 2026.\u003c/figcaption>\u003c/figure>\n",
"innerContent": [
"\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg\" alt=\"\" class=\"wp-image-12069061\" />\u003cfigcaption class=\"wp-element-caption\">State Sen. Scott Wiener, a candidate for California’s 11th Congressional District, participates in a forum with other candidates at UC Law San Francisco on Jan. 7, 2026.\u003c/figcaption>\u003c/figure>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>AI, he said, is “likely the most powerful technology in human history, and we need to make sure that we are both understanding the risks and taking them seriously, so that we can get ahead of them.” \u003c/p>\n",
"innerContent": [
"\n\u003cp>AI, he said, is “likely the most powerful technology in human history, and we need to make sure that we are both understanding the risks and taking them seriously, so that we can get ahead of them.” \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>Calvin likened the Hugging Face hack to the cyclospora outbreak linked to, but not confirmed as originating from, \u003ca href=\"https://www.kqed.org/news/12091770/california-based-taylor-farms-recalls-lettuce-shipped-to-27-states-over-cyclospora-risk\">Taylor Farms\u003c/a>. \u003c/p>\n",
"innerContent": [
"\n\u003cp>Calvin likened the Hugging Face hack to the cyclospora outbreak linked to, but not confirmed as originating from, \u003ca href=\"https://www.kqed.org/news/12091770/california-based-taylor-farms-recalls-lettuce-shipped-to-27-states-over-cyclospora-risk\">Taylor Farms\u003c/a>. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>That California-based company has said it’s removing its lettuce indefinitely while the FDA investigation continues. \u003c/p>\n",
"innerContent": [
"\n\u003cp>That California-based company has said it’s removing its lettuce indefinitely while the FDA investigation continues. \u003c/p>\n"
],
"innerBlocks": []
},
{
"blockName": "kqed/ad",
"attrs": {
"format": "floatright"
},
"innerBlocks": [],
"innerHTML": "",
"innerContent": []
},
{
"blockName": "core/paragraph",
"attrs": [],
"innerHTML": "\n\u003cp>“And OpenAI is like, ‘Maybe it’ll happen again. Maybe it’ll get worse. We don’t really know,“ Calvin said. “Your AI, again, hacked out of its box and hacked into another company, and you’re saying that you don’t know how to stop it from doing that again? That seems pretty nuts,” Calvin said.\u003c/p>\n",
"innerContent": [
"\n\u003cp>“And OpenAI is like, ‘Maybe it’ll happen again. Maybe it’ll get worse. We don’t really know,“ Calvin said. “Your AI, again, hacked out of its box and hacked into another company, and you’re saying that you don’t know how to stop it from doing that again? That seems pretty nuts,” Calvin said.\u003c/p>\n"
],
"innerBlocks": []
}
],
"excerpt": "California’s new frontier AI law doesn't require OpenAI to report its agents’ Hugging Face Hack. How often does this happen elsewhere?",
"status": "publish",
"parent": 0,
"modified": 1784850439,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 38,
"wordCount": 1697
},
"headData": {
"title": "How OpenAI’s Models Escaped Their Sandbox and Slipped Past California's AI Law | KQED",
"description": "California’s new frontier AI law doesn't require OpenAI to report its agents’ Hugging Face Hack. How often does this happen elsewhere?",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "How OpenAI’s Models Escaped Their Sandbox and Slipped Past California's AI Law",
"datePublished": "2026-07-23T16:47:00-07:00",
"dateModified": "2026-07-23T16:47:19-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"audioUrl": "https://traffic.omny.fm/d/clips/0af137ef-751e-4b19-a055-aaef00d2d578/ffca7e9f-6831-41c5-bcaf-aaef00f5a073/2059a50c-ed98-4446-bea8-b490013565f3/audio.mp3",
"sticky": false,
"nprStoryId": "kqed-12092162",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12092162/how-openais-models-escaped-their-sandbox-and-slipped-past-californias-ai-law",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>\u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI\u003c/a>\u003ca href=\"https://openai.com/index/hugging-face-model-evaluation-security-incident/\"> this week said\u003c/a> more than one of its models — three, according to \u003ca href=\"https://www.bloomberg.com/news/articles/2026-07-23/openai-models-lurked-in-hugging-face-system-for-hours-undetected?srnd=homepage-americas\">\u003cem>Bloomberg\u003c/em>\u003c/a> — attacked Hugging Face, a popular repository for open-source AI tools — marking one of the first publicly disclosed cases of frontier AI models autonomously cyberattacking another company. \u003c/p>\n\n\n\n\u003cp>The disclosure came days after Hugging Face \u003ca href=\"https://huggingface.co/blog/security-incident-july-2026\">revealed\u003c/a> it was breached in a hack by a “malicious dataset,” which intruded into the AI company’s software to run code. The campaign executed a flurry of more than 17,000 automated actions in a matter of hours, the post said. It’s still unclear if Hugging Face customer or partner data was taken. \u003c/p>\n\n\n\n\u003cp>“This matches the ‘agentic attacker’ scenario the industry has been forecasting,” Hugging Face wrote on July 16.\u003c/p>\n\n\n\n\u003cp>OpenAI said an investigation showed the incident was driven by a combination of models — “including GPT‑5.6 Sol and an even more capable pre-release model.” The company said it’s working with Hugging Face to produce a more thorough report of what happened.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/05/AP24134775174210-scaled-e1770337042768.jpg\" alt=\"\" class=\"wp-image-11985952\">\u003cfigcaption class=\"wp-element-caption\">The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston, Massachusetts. (Michael Dwyer/Associated Press)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>OpenAI’s announcement came one day after the San Francisco-based frontier AI developer disclosed a separate incident in which it paused another pre-release model that had escaped what the industry calls “a sandbox” — an isolated environment with no path to the open internet, except for a single internal service that fetches software packages — and \u003ca href=\"https://www.wired.com/story/openai-models-escaped-containment-and-hacked-huggingface/\">posted\u003c/a> to GitHub. \u003c/p>\n\n\n\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>Last April, a researcher challenged an early version of Anthropic’s Mythos model to break out of an isolated system and message back. Mythos did, and then \u003ca href=\"https://www.mondaymomentum.io/p/the-ai-that-broke-out-of-its-cage\">went further\u003c/a>, building a multistep process to reach the wider internet. \u003c/p>\n\n\n\n\u003cp>California’s \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">frontier AI law\u003c/a>, the first of its kind in the country, obliges the largest developers to tell the state about critical safety incidents, but only ones that kill or injure someone or that cause catastrophic harm.\u003c/p>\n\n\n\n\n\n\u003cp>It expressly excludes the kind of safety evaluation OpenAI was running when its models got loose. Which leaves a question nobody in Sacramento or Washington can answer: how often have frontier models in development gone rogue at companies that decided not to mention it? \u003c/p>\n\n\n\n\u003cp>“This was one of the topics I was focused on when I worked at OpenAI: what happens when AI models become good enough to reward-hack their way through the infrastructure they use to accomplish their objectives?” Ariel Herbert-Voss, OpenAI’s first security hire, wrote KQED. \u003c/p>\n\n\n\n\u003cp>“It’s a really difficult problem to defend against and is one of the reasons I left to start my own company,” Herbert-Voss, now the founder of an automated offensive cybersecurity company called RunSybil, said.\u003c/p>\n\n\n\n\u003cp>“It’s a really remarkable and, I think, fair to say scary event,“ said Nathan Calvin, general counsel at safety advocate EncodeAI. \u003c/p>\n\n\n\n\u003cp>Scientists and AI researchers have warned for years of the possibility that AI systems could go rogue and do unpredictable and unintended things. AI developers commonly test safety with so-called “\u003ca href=\"https://www.paloaltonetworks.com/cyberpedia/what-is-ai-red-teaming\">red teams\u003c/a>” that try to break or trick machine learning models, like chatbots, to find hidden flaws. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"1980\" height=\"1320\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg\" alt=\"\" class=\"wp-image-12092368\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862.jpg 1980w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/07/GettyImages-1849294862-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 1980px) 100vw, 1980px\">\u003cfigcaption class=\"wp-element-caption\">The Hugging Face logo is seen displayed on a mobile phone screen with the AI revolution symbol in the background. (Idrees Abbas/SOPA Images/LightRocket via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>“In certain circumstances, AI will try to break out of their sandboxes or otherwise do things that were not what the creator intended,” Calvin said. “But this really is the first very big example of that happening, at scale, with a really highly capable AI model in a way that actually harmed a third party.”\u003c/p>\n\n\n\n\u003cp>According to OpenAI’s account, the company was running an internal evaluation designed to measure how good its models are at hacking. To get a true reading, OpenAI deliberately switched off the safety filters that normally stop its models from carrying out cyberattacks.\u003c/p>\n\n\n\n\u003cp>The models were confined to a so-called “sandbox,” an isolated environment with no route to the open internet except OpenAI’s one internal service that fetches software libraries from public repositories. The models found a previously unknown flaw in that service, exploited it, escalated their privileges and moved from machine to machine inside OpenAI’s research network until they reached one with internet access. \u003c/p>\n\n\n\n\u003cp>Once the models broke out onto the open internet, according to OpenAI, they reasoned that Hugging Face might be storing the answers to the test they were trying to pass. Hugging Face’s credentials and production database aren’t on the public internet any more than a bank’s vault sits on a sidewalk, but the models broke into Hugging Face using stolen credentials, among other things.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg\" alt=\"\" class=\"wp-image-12062600\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01.jpeg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-160x107.jpeg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/10/111623_Sam-Altman_AP_CM_01-1536x1024.jpeg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">OpenAI restructured to a for-profit company in a move tied to an agreement with California Attorney General Rob Bonta. Company CEO Sam Altman participates in a discussion in San Francisco on Nov. 16, 2023. (Photo by Eric Risberg, AP Photo)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In short: The models were told to hack. They were not told to leave the building or to break into another company’s servers.\u003c/p>\n\n\n\n\u003cp>Hugging Face detected the intrusion and shut it down on its own, days before OpenAI connected the attack to its own testing. \u003c/p>\n\n\n\n\u003cp>“It’s quite mind-blowing that all of this happened autonomously!“ Hugging Face Chief Executive Clement Delangue wrote on social media platform \u003ca href=\"https://x.com/ClementDelangue/status/2079670308156645882\">X\u003c/a>.\u003c/p>\n\n\n\n\u003cp>In its blog post, OpenAI wrote, “This incident points to the need to further strengthen our model’s alignment, cyber protections during evaluation time and monitoring during internal testing.”\u003c/p>\n\n\n\n\u003cp>The Future of Life Institute — a nonprofit which issues a biannual risk assessment of nine leading AI companies — recently warned that many of the companies building frontier models are quietly walking back safety commitments. \u003c/p>\n\n\n\n\u003cp>According to the group’s \u003ca href=\"https://futureoflife.org/ai-safety-index-summer-2026/\">most recent\u003c/a> AI Safety Index, Anthropic, OpenAI, Google DeepMind and Meta all weakened or abandoned promises to pause development if certain red lines were approached, even while publicly suggesting they were amenable. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1334\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" class=\"wp-image-12058035\">\u003cfigcaption class=\"wp-element-caption\">Close-up of phone screen displaying Anthropic Claude, a Large Language Model (LLM) powered generative artificial intelligence chatbot in Lafayette, California, on June 27, 2024. (Smith Collection/Gado via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>In an email, KQED asked Hamza Chaudhry, who leads AI and national security work at Future of Life, to imagine it was not OpenAI’s software gone rogue but a foreign state, intentionally executing an unauthorized intrusion into a private company’s production infrastructure, exploiting cybersecurity vulnerabilities, stealing live credentials and accessing a production database. \u003c/p>\n\n\n\n\u003cp>“We would likely call this a dangerous act of cyber-espionage,” he wrote — a likely criminal violation of the Computer Fraud and Abuse Act, which “would draw a threat group designation and eventually an indictment or sanctions.”\u003c/p>\n\n\n\n\u003cp>OpenAI did not respond to KQED’s request for comment, but a spokesperson told \u003cem>Bloomberg \u003c/em>the company communicated with law enforcement and other government authorities about the incident.\u003c/p>\n\n\n\n\u003cp>The fact that Google DeepMind, Meta, Mistral, xAI and Chinese AI model developers haven’t revealed similar events doesn’t necessarily mean they haven’t happened. OpenAI and Anthropic are the only two frontier model developers that have publicly disclosed containment failures. \u003c/p>\n\n\n\n\u003cp>There is no mandatory public disclosure regime in force, like the one in California that requires \u003ca href=\"https://www.kqed.org/news/11992954/cybersecurity-expert-shares-tips-for-dublin-based-patelco-credit-union-customers-after-ransomware-attack\">hacked companies to reveal\u003c/a> there’s been a data breach.\u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg\" alt=\"\" class=\"wp-image-12036125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-800x533.jpg 800w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1020x680.jpg 1020w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1536x1024.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2025/04/MetaGetty2-1920x1280.jpg 1920w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">The Meta, Facebook, Instagram, WhatsApp, Messenger and Threads logos are screened on a mobile phone on Jan. 25, 2025. (Beata Zawrzel/NurPhoto via Getty Images)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>Congressional candidate and state Sen. Scott Wiener wrote two AI frontier model safety bills. Gov. Gavin Newsom \u003ca href=\"https://www.kqed.org/news/12007323/can-california-still-lead-on-ai-regulation-following-newsoms-veto-of-ai-safety-bill\">vetoed\u003c/a> Wiener’s first effort, arguing in his \u003ca href=\"https://www.gov.ca.gov/wp-content/uploads/2024/09/SB-1047-Veto-Message.pdf\">veto message\u003c/a> that “By focusing only on the most expensive and large-scale models, SB 1047 establishes a regulatory framework that could give the public a false sense of security about controlling this fast-moving technology.” \u003c/p>\n\n\n\n\u003cp>The next year, Newsom signed Wiener’s second at-bat, but only after the bill was softened to \u003ca href=\"https://www.kqed.org/news/12058013/newsom-signs-california-ai-transparency-bill-tailored-to-meet-tech-industry-tastes\">overcome industry pushback\u003c/a>. \u003c/p>\n\n\n\n\u003cp>The \u003ca href=\"https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202520260SB53\">law\u003c/a>, which took effect Jan. 1, requires developers of the most powerful AI models to notify the Governor’s Office of Emergency Services of any “critical safety incident“ within 15 days of discovering it — which is not the same as reporting the incident to the public. \u003c/p>\n\n\n\n\u003cp>Wiener told KQED he still thinks the law is strong. \u003c/p>\n\n\n\n\u003cp>“We worked very hard with the governor to produce a bill that is meaningful and impactful, and that he would sign,“ he said, adding he doesn’t consider the work done, because AI continues to evolve at a rapid pace. \u003c/p>\n\n\n\n\u003cfigure class=\"wp-block-image size-full\">\u003cimg loading=\"lazy\" decoding=\"async\" width=\"2000\" height=\"1333\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg\" alt=\"\" class=\"wp-image-12069061\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/01/260107-SFCongressionalCandidateForum-13-BL_qed-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-element-caption\">State Sen. Scott Wiener, a candidate for California’s 11th Congressional District, participates in a forum with other candidates at UC Law San Francisco on Jan. 7, 2026. (Beth LaBerge/KQED)\u003c/figcaption>\u003c/figure>\n\n\n\n\u003cp>AI, he said, is “likely the most powerful technology in human history, and we need to make sure that we are both understanding the risks and taking them seriously, so that we can get ahead of them.” \u003c/p>\n\n\n\n\u003cp>Calvin likened the Hugging Face hack to the cyclospora outbreak linked to, but not confirmed as originating from, \u003ca href=\"https://www.kqed.org/news/12091770/california-based-taylor-farms-recalls-lettuce-shipped-to-27-states-over-cyclospora-risk\">Taylor Farms\u003c/a>. \u003c/p>\n\n\n\n\u003cp>That California-based company has said it’s removing its lettuce indefinitely while the FDA investigation continues. \u003c/p>\n\n\n\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "floatright"
},
"numeric": [
"floatright"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>“And OpenAI is like, ‘Maybe it’ll happen again. Maybe it’ll get worse. We don’t really know,“ Calvin said. “Your AI, again, hacked out of its box and hacked into another company, and you’re saying that you don’t know how to stop it from doing that again? That seems pretty nuts,” Calvin said.\u003c/p>\n\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12092162/how-openais-models-escaped-their-sandbox-and-slipped-past-californias-ai-law",
"authors": [
"251"
],
"categories": [
"news_28250",
"news_8",
"news_248"
],
"tags": [
"news_25184",
"news_34755",
"news_32668",
"news_27626",
"news_35758",
"news_38",
"news_34586",
"news_1631"
],
"featImg": "news_12060375",
"label": "news"
},
"news_12087559": {
"type": "posts",
"id": "news_12087559",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12087559",
"score": null,
"sort": [
1781546455000
]
},
"guestAuthors": [],
"slug": "california-admits-using-high-risk-ai-including-systems-it-failed-to-report-last-year",
"title": "California Admits Using High-Risk AI — Including Systems It Failed to Report Last Year",
"publishDate": 1781546455,
"format": "standard",
"headTitle": "California Admits Using High-Risk AI — Including Systems It Failed to Report Last Year | KQED",
"labelTerm": {},
"content": "\u003cp>\u003c!-- Creative Commons Attribution-NonCommercial-NoDerivatives https://creativecommons.org/licenses/by-nc-nd/4.0/ -->\u003c/p>\n\u003cp>\u003cem>This story was originally published by \u003ca href=\"https://calmatters.org/\">CalMatters\u003c/a>. \u003ca href=\"https://calmatters.org/subscribe-to-calmatters/\">Sign up\u003c/a> for their newsletters.\u003c/em>\u003c/p>\n\u003cp>A year ago, California officials had to report under a new state law how they used automated systems to make important decisions about people’s lives.\u003c/p>\n\u003cp>They said they never did — a startling answer for a number of reasons, sources \u003ca href=\"https://calmatters.org/economy/technology/2025/05/california-somehow-finds-no-ai-risks/\">told \u003cem>CalMatters\u003c/em> at the time\u003c/a>, including that there were several prominent examples to the contrary.\u003c/p>\n\u003cp>Now, the state has issued a more expansive answer: It is currently using six automated systems to make consequential decisions about the lives of Californians.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>The systems are used to do things like:\u003c/p>\n\u003cul>\n\u003cli>Predict whether incarcerated people will reoffend\u003c/li>\n\u003cli>Evaluate whether unemployment claims are fraudulent\u003c/li>\n\u003cli>Remotely administer exams for California State University students\u003c/li>\n\u003cli>Detect when college students use generative AI to write assignments\u003c/li>\n\u003c/ul>\n\u003cp>That’s according to a \u003ca href=\"https://calmatters.org/wp-content/uploads/2026/06/High-Risk-ADS-Report-for-Program-Year-2025.pdf\">report released Friday\u003c/a> by the state’s technology department. The report is required \u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202320240ab302\">under a 2023 law mandating that\u003c/a> that state agencies annually disclose their use of “high-risk automated decision systems,” which the law defines as systems “used to assist or replace human discretionary decisions that have a legal or similarly significant effect, including decisions that materially impact access to, or approval for, housing or accommodations, education, employment, credit, healthcare, and criminal justice.”\u003c/p>\n\u003cp>The law was pushed by civil rights, privacy and civil liberties groups concerned about harms from AI-like systems. Numerous such systems have been shown to produce results biased against marginalized groups, including those used for \u003ca href=\"https://venturebeat.com/technology/examsofts-remote-bar-exam-sparks-privacy-and-facial-recognition-concerns\">high-stakes testing\u003c/a>, \u003ca href=\"https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing\">predicting recidivism\u003c/a>, and \u003ca href=\"https://themarkup.org/machine-learning/2023/08/14/ai-detection-tools-falsely-accuse-international-students-of-cheating\">detecting AI-generated texts\u003c/a>.\u003c/p>\n\u003cfigure id=\"attachment_12058035\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12058035 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">Anthropic’s AI model, Claude, a Large Language Model (LLM) is displayed on an iPhone in Lafayette, California, on June 27, 2024. \u003ccite>(Smith Collection/Gado via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>\u003ca href=\"https://calmatters.org/economy/technology/2025/05/california-somehow-finds-no-ai-risks/\">\u003cem>CalMatters\u003c/em> flagged last year’s report\u003c/a> as surprising, noting that the state corrections department had reported using software to predict post-release behavior and that the employment department used a fraud detection system that paused benefits for 600,000 Californians between Christmas and New Year’s in 2020, \u003ca href=\"https://lao.ca.gov/Publications/Report/4542\">according to a Legislative Analyst’s Office report\u003c/a>.\u003c/p>\n\u003cp>Though the report names six high-risk systems in use today, state agencies have used some for several years now. Those include COMPAS, which has been used by the California Department of Corrections and Rehabilitation to assign recidivism scores to inmates for at \u003ca href=\"https://www.cdcr.ca.gov/research/wp-content/uploads/sites/174/2023/06/Recidivism-Report-for-Offenders-Released-in-FY-2011-12.pdf\">least a decade\u003c/a>.\u003c/p>\n\u003cp>The technology department said in the report that it found more systems for its report this year because it evaluated responses from state agencies more thoroughly, including by meeting with agencies and questioning them about their systems.[aside postID=news_12087201 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CollegeGraduationGetty.jpg']In addition to the six high-risk systems, the department’s report disclosed an additional six systems initially flagged as high risk but later determined not to be. One was AI used for legislative bill analysis by the California Department of Finance.\u003c/p>\n\u003cp>The report also notes two high-risk systems that are not currently in use: the Department of Cannabis Control is developing artificial intelligence to analyze whether marijuana packaging violates a law against appealing to children, and California State University discontinued use of a language model for reviewing job applications.\u003c/p>\n\u003cp>Results of the second annual survey come after cities like San José and San Francisco released their first AI inventories in recent months.\u003c/p>\n\u003cp>They also come at a time when California-based AI companies like Anthropic and OpenAI are going public and seeking government contracts. Americans are \u003ca href=\"https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence/\">split on whether they trust AI\u003c/a>, and surveys last year by \u003ca href=\"https://techequity.us/press_release/californians-are-more-concerned-than-excited-by-ai/\">TechEquity\u003c/a> and \u003ca href=\"https://carnegieendowment.org/research/2025/10/carnegie-california-ai-survey\">Carnegie California found\u003c/a> that the majority of Californians want safety over innovation. A Gallup poll to evaluate the opinions of Americans found similar results.\u003c/p>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202520260sb1248\">Senate Bill 1248\u003c/a>, a bill that would have prohibited state employees from using automated decision systems as the sole basis for decision-making, was killed last month in the state’s \u003ca href=\"https://calmatters.org/politics/2026/05/suspense-file-senate-assembly/\">rapid-fire appropriations process\u003c/a>.\u003c/p>\n\u003ch2>What’s missing\u003c/h2>\n\u003cp>While the newly released report shares more information than last year’s, several questions remain about the state’s use of artificial intelligence and other automated systems.\u003c/p>\n\u003cp>The report does not include generative AI pilot projects underway with support from the governor’s office to do things like help businesses file taxes, support state employees who work on homelessness and an AI assistant named Poppy that uses language models like Anthropic’s Claude to do things like draft documents, research policy, or build custom AI tools, according to \u003ca href=\"https://www.genai.ca.gov/poppy/\">a state website\u003c/a>. The website says that 67 state departments provided input during the pilot phase, and the \u003ca href=\"https://www.genai.ca.gov/poppy/\">statewide rollout of Poppy begins next month\u003c/a>.\u003c/p>\n\u003cfigure id=\"attachment_11989313\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11989313\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-scaled-e1760733694503.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">The OpenAI ChatGPT logo. \u003ccite>(Jaap Arriens/NurPhoto via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>A \u003ca href=\"https://www.npr.org/2026/05/25/nx-s1-5772820/artificial-intelligence-education-technology-california-state-university\">California State University contract\u003c/a> with OpenAI to provide a version of ChatGPT is also not mentioned, though surveys of AI use in educational settings have found that \u003ca href=\"https://calmatters.org/economy/technology/2026/02/ai-images-scandalized-a-california-elementary-school-now-the-state-is-pushing-new-safeguards/\">the technology can do more harm than good\u003c/a>.\u003c/p>\n\u003cp>The 2023 law mandating the annual high-risk systems report excludes reporting by a number of state agencies, including the judicial branch and the University of California college system. \u003ca href=\"https://calmatters.org/economy/technology/2026/05/ai-los-angeles-riverside-courts/\">Reporting by \u003cem>CalMatters\u003c/em> last month\u003c/a> found that a majority of the roughly 60 courts that operate statewide have adopted generative AI use policies. Courts in Los Angeles and Riverside counties have \u003ca href=\"https://calmatters.org/economy/technology/2026/05/ai-los-angeles-riverside-courts/\">begun testing\u003c/a> an AI tool to act as a clerk, drafting orders and producing research memos.\u003c/p>\n\u003cp>\u003cem>CalMatters\u003c/em> is compiling an inventory of automated decision-making systems in use by state and local agencies throughout California in order to provide transparency into how governments are using decision-making systems and AI. Know about an AI system in use by a state or local agency? Email \u003ca href=\"mailto:khari@calmatters.org\">khari@calmatters.org\u003c/a>.\u003c/p>\n\u003cp>\u003cem>This article was \u003ca href=\"https://calmatters.org/politics/2026/06/california-admits-government-ai-risk-after-denying/\">originally published on CalMatters\u003c/a> and was republished under the \u003ca href=\"https://creativecommons.org/licenses/by-nc-nd/4.0/\">Creative Commons Attribution-NonCommercial-NoDerivatives\u003c/a> license.\u003c/em>\u003c/p>\n\u003cp>\u003c/p>\n",
"blocks": [],
"excerpt": "State officials have found they are using six high-risk AI-like systems. One year ago, they reported using zero.",
"status": "publish",
"parent": 0,
"modified": 1781542932,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 23,
"wordCount": 978
},
"headData": {
"title": "California Admits Using High-Risk AI — Including Systems It Failed to Report Last Year | KQED",
"description": "State officials have found they are using six high-risk AI-like systems. One year ago, they reported using zero.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "California Admits Using High-Risk AI — Including Systems It Failed to Report Last Year",
"datePublished": "2026-06-15T11:00:55-07:00",
"dateModified": "2026-06-15T10:02:12-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"source": "CalMatters",
"sourceUrl": "https://calmatters.org/",
"sticky": false,
"nprByline": "Khari Johnson, CalMatters",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"showOnAuthorArchivePages": "No",
"articleAge": "0",
"path": "/news/12087559/california-admits-using-high-risk-ai-including-systems-it-failed-to-report-last-year",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>\u003c!-- Creative Commons Attribution-NonCommercial-NoDerivatives https://creativecommons.org/licenses/by-nc-nd/4.0/ -->\u003c/p>\n\u003cp>\u003cem>This story was originally published by \u003ca href=\"https://calmatters.org/\">CalMatters\u003c/a>. \u003ca href=\"https://calmatters.org/subscribe-to-calmatters/\">Sign up\u003c/a> for their newsletters.\u003c/em>\u003c/p>\n\u003cp>A year ago, California officials had to report under a new state law how they used automated systems to make important decisions about people’s lives.\u003c/p>\n\u003cp>They said they never did — a startling answer for a number of reasons, sources \u003ca href=\"https://calmatters.org/economy/technology/2025/05/california-somehow-finds-no-ai-risks/\">told \u003cem>CalMatters\u003c/em> at the time\u003c/a>, including that there were several prominent examples to the contrary.\u003c/p>\n\u003cp>Now, the state has issued a more expansive answer: It is currently using six automated systems to make consequential decisions about the lives of Californians.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>The systems are used to do things like:\u003c/p>\n\u003cul>\n\u003cli>Predict whether incarcerated people will reoffend\u003c/li>\n\u003cli>Evaluate whether unemployment claims are fraudulent\u003c/li>\n\u003cli>Remotely administer exams for California State University students\u003c/li>\n\u003cli>Detect when college students use generative AI to write assignments\u003c/li>\n\u003c/ul>\n\u003cp>That’s according to a \u003ca href=\"https://calmatters.org/wp-content/uploads/2026/06/High-Risk-ADS-Report-for-Program-Year-2025.pdf\">report released Friday\u003c/a> by the state’s technology department. The report is required \u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202320240ab302\">under a 2023 law mandating that\u003c/a> that state agencies annually disclose their use of “high-risk automated decision systems,” which the law defines as systems “used to assist or replace human discretionary decisions that have a legal or similarly significant effect, including decisions that materially impact access to, or approval for, housing or accommodations, education, employment, credit, healthcare, and criminal justice.”\u003c/p>\n\u003cp>The law was pushed by civil rights, privacy and civil liberties groups concerned about harms from AI-like systems. Numerous such systems have been shown to produce results biased against marginalized groups, including those used for \u003ca href=\"https://venturebeat.com/technology/examsofts-remote-bar-exam-sparks-privacy-and-facial-recognition-concerns\">high-stakes testing\u003c/a>, \u003ca href=\"https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing\">predicting recidivism\u003c/a>, and \u003ca href=\"https://themarkup.org/machine-learning/2023/08/14/ai-detection-tools-falsely-accuse-international-students-of-cheating\">detecting AI-generated texts\u003c/a>.\u003c/p>\n\u003cfigure id=\"attachment_12058035\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12058035 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2025/09/GettyImages-2159671948-scaled-e1781542152687.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">Anthropic’s AI model, Claude, a Large Language Model (LLM) is displayed on an iPhone in Lafayette, California, on June 27, 2024. \u003ccite>(Smith Collection/Gado via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>\u003ca href=\"https://calmatters.org/economy/technology/2025/05/california-somehow-finds-no-ai-risks/\">\u003cem>CalMatters\u003c/em> flagged last year’s report\u003c/a> as surprising, noting that the state corrections department had reported using software to predict post-release behavior and that the employment department used a fraud detection system that paused benefits for 600,000 Californians between Christmas and New Year’s in 2020, \u003ca href=\"https://lao.ca.gov/Publications/Report/4542\">according to a Legislative Analyst’s Office report\u003c/a>.\u003c/p>\n\u003cp>Though the report names six high-risk systems in use today, state agencies have used some for several years now. Those include COMPAS, which has been used by the California Department of Corrections and Rehabilitation to assign recidivism scores to inmates for at \u003ca href=\"https://www.cdcr.ca.gov/research/wp-content/uploads/sites/174/2023/06/Recidivism-Report-for-Offenders-Released-in-FY-2011-12.pdf\">least a decade\u003c/a>.\u003c/p>\n\u003cp>The technology department said in the report that it found more systems for its report this year because it evaluated responses from state agencies more thoroughly, including by meeting with agencies and questioning them about their systems.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12087201",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/06/CollegeGraduationGetty.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>In addition to the six high-risk systems, the department’s report disclosed an additional six systems initially flagged as high risk but later determined not to be. One was AI used for legislative bill analysis by the California Department of Finance.\u003c/p>\n\u003cp>The report also notes two high-risk systems that are not currently in use: the Department of Cannabis Control is developing artificial intelligence to analyze whether marijuana packaging violates a law against appealing to children, and California State University discontinued use of a language model for reviewing job applications.\u003c/p>\n\u003cp>Results of the second annual survey come after cities like San José and San Francisco released their first AI inventories in recent months.\u003c/p>\n\u003cp>They also come at a time when California-based AI companies like Anthropic and OpenAI are going public and seeking government contracts. Americans are \u003ca href=\"https://www.pewresearch.org/short-reads/2026/03/12/key-findings-about-how-americans-view-artificial-intelligence/\">split on whether they trust AI\u003c/a>, and surveys last year by \u003ca href=\"https://techequity.us/press_release/californians-are-more-concerned-than-excited-by-ai/\">TechEquity\u003c/a> and \u003ca href=\"https://carnegieendowment.org/research/2025/10/carnegie-california-ai-survey\">Carnegie California found\u003c/a> that the majority of Californians want safety over innovation. A Gallup poll to evaluate the opinions of Americans found similar results.\u003c/p>\n\u003cp>\u003ca href=\"https://calmatters.digitaldemocracy.org/bills/ca_202520260sb1248\">Senate Bill 1248\u003c/a>, a bill that would have prohibited state employees from using automated decision systems as the sole basis for decision-making, was killed last month in the state’s \u003ca href=\"https://calmatters.org/politics/2026/05/suspense-file-senate-assembly/\">rapid-fire appropriations process\u003c/a>.\u003c/p>\n\u003ch2>What’s missing\u003c/h2>\n\u003cp>While the newly released report shares more information than last year’s, several questions remain about the state’s use of artificial intelligence and other automated systems.\u003c/p>\n\u003cp>The report does not include generative AI pilot projects underway with support from the governor’s office to do things like help businesses file taxes, support state employees who work on homelessness and an AI assistant named Poppy that uses language models like Anthropic’s Claude to do things like draft documents, research policy, or build custom AI tools, according to \u003ca href=\"https://www.genai.ca.gov/poppy/\">a state website\u003c/a>. The website says that 67 state departments provided input during the pilot phase, and the \u003ca href=\"https://www.genai.ca.gov/poppy/\">statewide rollout of Poppy begins next month\u003c/a>.\u003c/p>\n\u003cfigure id=\"attachment_11989313\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-11989313\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-scaled-e1760733694503.jpg\" alt=\"\" width=\"2000\" height=\"1334\">\u003cfigcaption class=\"wp-caption-text\">The OpenAI ChatGPT logo. \u003ccite>(Jaap Arriens/NurPhoto via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>A \u003ca href=\"https://www.npr.org/2026/05/25/nx-s1-5772820/artificial-intelligence-education-technology-california-state-university\">California State University contract\u003c/a> with OpenAI to provide a version of ChatGPT is also not mentioned, though surveys of AI use in educational settings have found that \u003ca href=\"https://calmatters.org/economy/technology/2026/02/ai-images-scandalized-a-california-elementary-school-now-the-state-is-pushing-new-safeguards/\">the technology can do more harm than good\u003c/a>.\u003c/p>\n\u003cp>The 2023 law mandating the annual high-risk systems report excludes reporting by a number of state agencies, including the judicial branch and the University of California college system. \u003ca href=\"https://calmatters.org/economy/technology/2026/05/ai-los-angeles-riverside-courts/\">Reporting by \u003cem>CalMatters\u003c/em> last month\u003c/a> found that a majority of the roughly 60 courts that operate statewide have adopted generative AI use policies. Courts in Los Angeles and Riverside counties have \u003ca href=\"https://calmatters.org/economy/technology/2026/05/ai-los-angeles-riverside-courts/\">begun testing\u003c/a> an AI tool to act as a clerk, drafting orders and producing research memos.\u003c/p>\n\u003cp>\u003cem>CalMatters\u003c/em> is compiling an inventory of automated decision-making systems in use by state and local agencies throughout California in order to provide transparency into how governments are using decision-making systems and AI. Know about an AI system in use by a state or local agency? Email \u003ca href=\"mailto:khari@calmatters.org\">khari@calmatters.org\u003c/a>.\u003c/p>\n\u003cp>\u003cem>This article was \u003ca href=\"https://calmatters.org/politics/2026/06/california-admits-government-ai-risk-after-denying/\">originally published on CalMatters\u003c/a> and was republished under the \u003ca href=\"https://creativecommons.org/licenses/by-nc-nd/4.0/\">Creative Commons Attribution-NonCommercial-NoDerivatives\u003c/a> license.\u003c/em>\u003c/p>\n\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12087559/california-admits-using-high-risk-ai-including-systems-it-failed-to-report-last-year",
"authors": [
"byline_news_12087559"
],
"categories": [
"news_31795",
"news_8",
"news_248"
],
"tags": [
"news_25184",
"news_34755",
"news_18538",
"news_32668",
"news_20013",
"news_16",
"news_34586",
"news_1631"
],
"affiliates": [
"news_18481"
],
"featImg": "news_12087570",
"label": "source_news_12087559"
},
"news_12084062": {
"type": "posts",
"id": "news_12084062",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12084062",
"score": null,
"sort": [
1779131568000
]
},
"guestAuthors": [],
"slug": "federal-court-rules-against-elon-musk-in-his-bitter-feud-with-sam-altman",
"title": "Federal Court Rules Against Elon Musk in His Bitter Feud With Sam Altman",
"publishDate": 1779131568,
"format": "standard",
"headTitle": "Federal Court Rules Against Elon Musk in His Bitter Feud With Sam Altman | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>Elon Musk’s \u003ca href=\"https://www.kqed.org/news/12081290/how-to-unscramble-an-omelet-in-silicon-valley-the-musk-v-altman-trial-that-will-try\">lawsuit against his OpenAI co-founders\u003c/a> has been rejected by a federal judge in Oakland, who found his claims were outside the statute of limitations.\u003c/p>\n\u003cp>Musk, who helped form OpenAI as a nonprofit in 2015, had alleged that co-founders Sam Altman and Greg Brockman violated the company’s original nonprofit mission to create safe and open-source artificial intelligence in order to enrich themselves. An Oakland jury took just a few hours to declare that Musk’s claim came too late. U.S. District Judge Yvonne Gonzalez Rogers, who had the final say in the case, agreed with the jury’s advisory verdict.\u003c/p>\n\u003cp>“The finding of the jury confirms that what this lawsuit was, was a hypocritical attempt to sabotage a competitor and to overcome a long history of very bad predictions about what OpenAI has been and will become,” Altman’s lead counsel, William Savitt, told reporters outside the courthouse Monday.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>The verdict comes after a weekslong blockbuster trial in Silicon Valley, in which the Tesla CEO accused Altman and Brockman of \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">“stealing a charity\u003c/a>” as they built a more than $850 million company on the back of their nonprofit. Court documents and testimony from \u003ca href=\"https://www.kqed.org/news/12083224/former-openai-exec-calls-decision-to-remove-sam-altman-a-hail-mary-during-musk-trial\">a score of tech elites\u003c/a>, including Microsoft CEO Satya Nadella, shed light on the rise of OpenAI — as well as on \u003ca href=\"https://www.kqed.org/news/12083278/sam-altman-defends-himself-from-elon-musks-accusations-in-openai-trial\">the interpersonal strife\u003c/a> and \u003ca href=\"https://www.kqed.org/news/12083612/lawyers-for-elon-musk-and-sam-altman-make-their-final-case-in-openai-trial\">falling out between Altman and Musk\u003c/a>, who were once close friends.\u003c/p>\n\u003cp>Musk’s high-profile testimony in the case also raised questions over Altman’s trustworthiness and leadership as the company pursues artificial general intelligence, a superintelligent form of AI and a potential trillion-dollar initial public offering.\u003c/p>\n\u003cfigure id=\"attachment_12084143\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12084143\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">The verdict is read in the trial in which Elon Musk claimed that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 18, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Altman’s defense argued that OpenAI had to form a profit-generating arm to keep up with competitors as AI technology advanced. They said that prior to leaving OpenAI, Musk was amenable to creating a for-profit, which he wanted to control. When other executives refused to agree to his terms, he left the company.\u003c/p>\n\u003cp>But Monday’s verdict disregarded many of the trial’s revelations, and instead hinged on the timeline of Musk’s claims.\u003c/p>\n\u003cp>The jury agreed with OpenAI’s defense that Musk missed the statute of limitations to allege a breach of charitable trust. They also dismissed a claim that Microsoft, OpenAI’s largest investor, aided and abetted a breach of charitable trust.[aside postID=news_12083612 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty.jpg']Sarah Eddy, an attorney for OpenAI, noted in her closing argument that Musk departed the company in 2018, watched it build up a for-profit arm beginning in 2019 and made his final monetary contribution the year after that. Yet, he waited until 2024, after he’d launched a competing AI enterprise, to bring his suit.\u003c/p>\n\u003cp>She called the case a “textbook” example of why the statute of limitations exists, saying that when Musk made his last contribution and testified that he became suspicious of a breach of charitable trust in 2020, he “started the clock.” According to Eddy, Musk should have sued by 2022 at the latest.\u003c/p>\n\u003cp>Musk’s counsel, Marc Toberoff, said there was a strong basis for appeal based on the legal components, statute of limitations aside. Musk also wrote on X, which he owns, that he planned to file an appeal.\u003c/p>\n\u003cp>“There is no question to anyone following the case in detail that Altman & Brockman did in fact enrich themselves by stealing a charity. The only question is WHEN they did it!” he wrote. “Creating a precedent to loot charities is incredibly destructive to charitable giving in America.”\u003c/p>\n\u003cp>Outside the court, anti-AI protesters who have been present for much of the trial decried the decision.\u003c/p>\n\u003cp>“No matter who won, we all lost,” said Phoebe Thomas Sorgen, an activist with StopAI, which seeks to “disrupt the reckless development of destructive” AI tech, according to its website. “We all lost. Sam Altman won, but look at who he is and what he’s doing.”\u003c/p>\n\u003cp>\u003c/p>\n",
"blocks": [],
"excerpt": "A U.S. judge and jury in Oakland found that the Tesla CEO filed his lawsuit against his former OpenAI co-founders too late. ",
"status": "publish",
"parent": 0,
"modified": 1779143179,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 15,
"wordCount": 744
},
"headData": {
"title": "Federal Court Rules Against Elon Musk in His Bitter Feud With Sam Altman | KQED",
"description": "A U.S. judge and jury in Oakland found that the Tesla CEO filed his lawsuit against his former OpenAI co-founders too late. ",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Federal Court Rules Against Elon Musk in His Bitter Feud With Sam Altman",
"datePublished": "2026-05-18T12:12:48-07:00",
"dateModified": "2026-05-18T15:26:19-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"sticky": false,
"nprStoryId": "kqed-12084062",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12084062/federal-court-rules-against-elon-musk-in-his-bitter-feud-with-sam-altman",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>Elon Musk’s \u003ca href=\"https://www.kqed.org/news/12081290/how-to-unscramble-an-omelet-in-silicon-valley-the-musk-v-altman-trial-that-will-try\">lawsuit against his OpenAI co-founders\u003c/a> has been rejected by a federal judge in Oakland, who found his claims were outside the statute of limitations.\u003c/p>\n\u003cp>Musk, who helped form OpenAI as a nonprofit in 2015, had alleged that co-founders Sam Altman and Greg Brockman violated the company’s original nonprofit mission to create safe and open-source artificial intelligence in order to enrich themselves. An Oakland jury took just a few hours to declare that Musk’s claim came too late. U.S. District Judge Yvonne Gonzalez Rogers, who had the final say in the case, agreed with the jury’s advisory verdict.\u003c/p>\n\u003cp>“The finding of the jury confirms that what this lawsuit was, was a hypocritical attempt to sabotage a competitor and to overcome a long history of very bad predictions about what OpenAI has been and will become,” Altman’s lead counsel, William Savitt, told reporters outside the courthouse Monday.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>The verdict comes after a weekslong blockbuster trial in Silicon Valley, in which the Tesla CEO accused Altman and Brockman of \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">“stealing a charity\u003c/a>” as they built a more than $850 million company on the back of their nonprofit. Court documents and testimony from \u003ca href=\"https://www.kqed.org/news/12083224/former-openai-exec-calls-decision-to-remove-sam-altman-a-hail-mary-during-musk-trial\">a score of tech elites\u003c/a>, including Microsoft CEO Satya Nadella, shed light on the rise of OpenAI — as well as on \u003ca href=\"https://www.kqed.org/news/12083278/sam-altman-defends-himself-from-elon-musks-accusations-in-openai-trial\">the interpersonal strife\u003c/a> and \u003ca href=\"https://www.kqed.org/news/12083612/lawyers-for-elon-musk-and-sam-altman-make-their-final-case-in-openai-trial\">falling out between Altman and Musk\u003c/a>, who were once close friends.\u003c/p>\n\u003cp>Musk’s high-profile testimony in the case also raised questions over Altman’s trustworthiness and leadership as the company pursues artificial general intelligence, a superintelligent form of AI and a potential trillion-dollar initial public offering.\u003c/p>\n\u003cfigure id=\"attachment_12084143\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12084143\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260518-MUSK-ALTMAN-TRIAL-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">The verdict is read in the trial in which Elon Musk claimed that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 18, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Altman’s defense argued that OpenAI had to form a profit-generating arm to keep up with competitors as AI technology advanced. They said that prior to leaving OpenAI, Musk was amenable to creating a for-profit, which he wanted to control. When other executives refused to agree to his terms, he left the company.\u003c/p>\n\u003cp>But Monday’s verdict disregarded many of the trial’s revelations, and instead hinged on the timeline of Musk’s claims.\u003c/p>\n\u003cp>The jury agreed with OpenAI’s defense that Musk missed the statute of limitations to allege a breach of charitable trust. They also dismissed a claim that Microsoft, OpenAI’s largest investor, aided and abetted a breach of charitable trust.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12083612",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/OpenAILawyerGetty.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Sarah Eddy, an attorney for OpenAI, noted in her closing argument that Musk departed the company in 2018, watched it build up a for-profit arm beginning in 2019 and made his final monetary contribution the year after that. Yet, he waited until 2024, after he’d launched a competing AI enterprise, to bring his suit.\u003c/p>\n\u003cp>She called the case a “textbook” example of why the statute of limitations exists, saying that when Musk made his last contribution and testified that he became suspicious of a breach of charitable trust in 2020, he “started the clock.” According to Eddy, Musk should have sued by 2022 at the latest.\u003c/p>\n\u003cp>Musk’s counsel, Marc Toberoff, said there was a strong basis for appeal based on the legal components, statute of limitations aside. Musk also wrote on X, which he owns, that he planned to file an appeal.\u003c/p>\n\u003cp>“There is no question to anyone following the case in detail that Altman & Brockman did in fact enrich themselves by stealing a charity. The only question is WHEN they did it!” he wrote. “Creating a precedent to loot charities is incredibly destructive to charitable giving in America.”\u003c/p>\n\u003cp>Outside the court, anti-AI protesters who have been present for much of the trial decried the decision.\u003c/p>\n\u003cp>“No matter who won, we all lost,” said Phoebe Thomas Sorgen, an activist with StopAI, which seeks to “disrupt the reckless development of destructive” AI tech, according to its website. “We all lost. Sam Altman won, but look at who he is and what he’s doing.”\u003c/p>\n\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12084062/federal-court-rules-against-elon-musk-in-his-bitter-feud-with-sam-altman",
"authors": [
"11913",
"251"
],
"categories": [
"news_6188",
"news_28250",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_1386",
"news_32668",
"news_3897",
"news_34054",
"news_33542",
"news_33543",
"news_34586",
"news_1631"
],
"featImg": "news_12084126",
"label": "news"
},
"news_12083612": {
"type": "posts",
"id": "news_12083612",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12083612",
"score": null,
"sort": [
1778805082000
]
},
"guestAuthors": [],
"slug": "lawyers-for-elon-musk-and-sam-altman-make-their-final-case-in-openai-trial",
"title": "Lawyers for Elon Musk and Sam Altman Make Their Final Case in OpenAI Trial",
"publishDate": 1778805082,
"format": "standard",
"headTitle": "Lawyers for Elon Musk and Sam Altman Make Their Final Case in OpenAI Trial | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>Whether \u003ca href=\"https://www.kqed.org/news/tag/openai\">OpenAI\u003c/a> CEO \u003ca href=\"https://www.kqed.org/news/tag/sam-altman\">Sam Altman\u003c/a> and other executives betrayed their commitment to building a safe, open-source artificial intelligence, slighting billionaire Elon Musk in the process, will be decided by an Oakland jury and judge.\u003c/p>\n\u003cp>For weeks, the tech executives have sparred in federal court over whether the startup, first proposed by Altman to Musk as a sort of AI “Manhattan Project,” has abandoned its original mission to enrich itself. Musk, who provided $38 million in early funding, has accused his former OpenAI co-founders Altman and Greg Brockman of “stealing a charity.”\u003c/p>\n\u003cp>OpenAI’s executives, on the other hand, have said Musk only sued after he brought his own AI competitor, xAI, onto the market.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>During his closing statement, Musk’s lead counsel, Steven Molo, focused on Altman’s credibility. He asked the jury to consider hypothetically what they would do if they came upon a bridge, suspended 150 feet above a river, and built on Altman’s “version of the truth.”\u003c/p>\n\u003cp>“Would you walk across that bridge?” He asked. “I don’t think many people would.”\u003c/p>\n\u003cp>Molo said that in the early years of OpenAI, the intent was to create a technology “for the good of the world.” He pointed to Musk’s early fears of the dangers of artificial general intelligence, or AGI, an early mission statement that said OpenAI would not be constrained by a need to generate financial return and correspondence between Altman and Musk that expressed support by both of them for a nonprofit structure and safety-focused mission.\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081686\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“There was no disagreement over the core mission,” Molo said.\u003c/p>\n\u003cp>But, he said, since OpenAI launched a for-profit subsidiary in 2019 — after Musk departed — Altman and his fellow executives have treated the nonprofit as a “shell,” transferring intellectual property and the vast majority of employees to the for-profit arm of the company. In 2023, Molo continued, after OpenAI made a $10 billion deal with Microsoft, the company failed to prioritize safety, abandoned its commitment to open sourcing and “enriched investors and insiders.”\u003c/p>\n\u003cp>“They’re motivated by money: Microsoft and Altman,” Molo said.\u003c/p>\n\u003cp>Microsoft CEO Satya Nadella testified that the company had invested $13 billion and expects to see a return of about $92 billion. Molo also pointed out that other executives, including Brockman and founding OpenAI computer scientist Ilya Sutskever, testified to having billions in equity, despite not investing in the company.[aside postID=news_12083278 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED.jpg']Altman’s attorneys argued that Musk’s case was baseless: not only was Molo’s characterization false, but they argued, the larger issue is that Musk’s contributions to OpenAI — in the form of rent payments, Tesla Model 3 cars and $25 million in quarterly donations — were never accompanied by specific promises for their use.\u003c/p>\n\u003cp>“If the donations came with no strings attached, then Mr. Musk does not have a charitable trust to enforce,” Sarah Eddy, an attorney for OpenAI’s defendants, said.\u003c/p>\n\u003cp>She and Altman’s lead counsel, William Savitt, also spent much of their closing arguments painting Musk as not wanting to protect humanity from AGI, as he’s suggested, but wanting to be the one who controls it.\u003c/p>\n\u003cp>They allege Musk brought his suit after he tried to wrest control of a potential for-profit arm of OpenAI, and later absorb the organization into Tesla, in 2017. The executives had begun discussing a for-profit expansion that year to solicit more funding for top talent and “compute” to compete with other industry leaders.\u003c/p>\n\u003cp>Musk departed OpenAI in February 2018, after a falling-out with the other executives over those discussions.\u003c/p>\n\u003cfigure id=\"attachment_12075382\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12075382\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">A courtroom sketch of Elon Musk on the stand as he’s questioned by the plaintiff’s attorney, Aaron P. Arnzen, on March 4, 2026. Musk is accused of making false and misleading statements that drove down Twitter’s stock price before he bought the social media platform for $44 billion in 2022. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Shortly after, OpenAI decided to introduce a for-profit public benefit corporation. It has since become a $850 billion company, and is considering an initial public offering estimated at up to a trillion dollars.\u003c/p>\n\u003cp>OpenAI’s attorneys alleged that Musk saw OpenAI’s skyrocketing success and filed his suit to destroy a competitor in the field.\u003c/p>\n\u003cp>“The truth is that Mr. Musk wanted a for-profit AI, and he wanted to dominate it,” Eddy said.\u003c/p>\n\u003cp>The jury is set to begin deliberations on Monday. If they side with Musk, OpenAI and Microsoft could owe $150 billion in damages to be redirected to the nonprofit foundation, along with a court order dismantling OpenAI’s for-profit structure and removal of Altman and Brockman from their posts.\u003c/p>\n\u003cfigure id=\"attachment_12083616\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12083616\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne.jpg\" alt=\"\" width=\"2000\" height=\"1334\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne-1536x1025.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Judge Yvonne Gonzalez Rogers of the U.S. District Court for the Northern District of California. \u003ccite>(Courtesy of Daily Journal)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>The jury will not have the final say, though. In a rare, but not unprecedented, move, U.S. District Judge Yvonne Gonzalez Rogers will have the ultimate right to rule on the claims.\u003c/p>\n\u003cp>According to Charlie Bullock, a senior research fellow at the Institute for Law and AI, this is because most times, “equitable claims” — breach of charitable trust and unjust enrichment — which involve non-monetary remedies, are decided by a judge.\u003c/p>\n\u003cp>In this case, Gonzalez Rogers elected to have an advisory jury, and Bullock said that typically, judges choose to go along with their decision.\u003c/p>\n\u003cp>\u003c/p>\n",
"blocks": [],
"excerpt": "An Oakland judge and jury will decide whether the company behind ChatGPT betrayed its mission of developing a safer, less risky AI. ",
"status": "publish",
"parent": 0,
"modified": 1778813223,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 23,
"wordCount": 1022
},
"headData": {
"title": "Lawyers for Elon Musk and Sam Altman Make Their Final Case in OpenAI Trial | KQED",
"description": "An Oakland judge and jury will decide whether the company behind ChatGPT betrayed its mission of developing a safer, less risky AI. ",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Lawyers for Elon Musk and Sam Altman Make Their Final Case in OpenAI Trial",
"datePublished": "2026-05-14T17:31:22-07:00",
"dateModified": "2026-05-14T19:47:03-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"sticky": false,
"nprStoryId": "kqed-12083612",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12083612/lawyers-for-elon-musk-and-sam-altman-make-their-final-case-in-openai-trial",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>Whether \u003ca href=\"https://www.kqed.org/news/tag/openai\">OpenAI\u003c/a> CEO \u003ca href=\"https://www.kqed.org/news/tag/sam-altman\">Sam Altman\u003c/a> and other executives betrayed their commitment to building a safe, open-source artificial intelligence, slighting billionaire Elon Musk in the process, will be decided by an Oakland jury and judge.\u003c/p>\n\u003cp>For weeks, the tech executives have sparred in federal court over whether the startup, first proposed by Altman to Musk as a sort of AI “Manhattan Project,” has abandoned its original mission to enrich itself. Musk, who provided $38 million in early funding, has accused his former OpenAI co-founders Altman and Greg Brockman of “stealing a charity.”\u003c/p>\n\u003cp>OpenAI’s executives, on the other hand, have said Musk only sued after he brought his own AI competitor, xAI, onto the market.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>During his closing statement, Musk’s lead counsel, Steven Molo, focused on Altman’s credibility. He asked the jury to consider hypothetically what they would do if they came upon a bridge, suspended 150 feet above a river, and built on Altman’s “version of the truth.”\u003c/p>\n\u003cp>“Would you walk across that bridge?” He asked. “I don’t think many people would.”\u003c/p>\n\u003cp>Molo said that in the early years of OpenAI, the intent was to create a technology “for the good of the world.” He pointed to Musk’s early fears of the dangers of artificial general intelligence, or AGI, an early mission statement that said OpenAI would not be constrained by a need to generate financial return and correspondence between Altman and Musk that expressed support by both of them for a nonprofit structure and safety-focused mission.\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081686\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“There was no disagreement over the core mission,” Molo said.\u003c/p>\n\u003cp>But, he said, since OpenAI launched a for-profit subsidiary in 2019 — after Musk departed — Altman and his fellow executives have treated the nonprofit as a “shell,” transferring intellectual property and the vast majority of employees to the for-profit arm of the company. In 2023, Molo continued, after OpenAI made a $10 billion deal with Microsoft, the company failed to prioritize safety, abandoned its commitment to open sourcing and “enriched investors and insiders.”\u003c/p>\n\u003cp>“They’re motivated by money: Microsoft and Altman,” Molo said.\u003c/p>\n\u003cp>Microsoft CEO Satya Nadella testified that the company had invested $13 billion and expects to see a return of about $92 billion. Molo also pointed out that other executives, including Brockman and founding OpenAI computer scientist Ilya Sutskever, testified to having billions in equity, despite not investing in the company.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12083278",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-03-KQED.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Altman’s attorneys argued that Musk’s case was baseless: not only was Molo’s characterization false, but they argued, the larger issue is that Musk’s contributions to OpenAI — in the form of rent payments, Tesla Model 3 cars and $25 million in quarterly donations — were never accompanied by specific promises for their use.\u003c/p>\n\u003cp>“If the donations came with no strings attached, then Mr. Musk does not have a charitable trust to enforce,” Sarah Eddy, an attorney for OpenAI’s defendants, said.\u003c/p>\n\u003cp>She and Altman’s lead counsel, William Savitt, also spent much of their closing arguments painting Musk as not wanting to protect humanity from AGI, as he’s suggested, but wanting to be the one who controls it.\u003c/p>\n\u003cp>They allege Musk brought his suit after he tried to wrest control of a potential for-profit arm of OpenAI, and later absorb the organization into Tesla, in 2017. The executives had begun discussing a for-profit expansion that year to solicit more funding for top talent and “compute” to compete with other industry leaders.\u003c/p>\n\u003cp>Musk departed OpenAI in February 2018, after a falling-out with the other executives over those discussions.\u003c/p>\n\u003cfigure id=\"attachment_12075382\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12075382\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/03/260304-Elon-Musk-Trial-01-KQED-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">A courtroom sketch of Elon Musk on the stand as he’s questioned by the plaintiff’s attorney, Aaron P. Arnzen, on March 4, 2026. Musk is accused of making false and misleading statements that drove down Twitter’s stock price before he bought the social media platform for $44 billion in 2022. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Shortly after, OpenAI decided to introduce a for-profit public benefit corporation. It has since become a $850 billion company, and is considering an initial public offering estimated at up to a trillion dollars.\u003c/p>\n\u003cp>OpenAI’s attorneys alleged that Musk saw OpenAI’s skyrocketing success and filed his suit to destroy a competitor in the field.\u003c/p>\n\u003cp>“The truth is that Mr. Musk wanted a for-profit AI, and he wanted to dominate it,” Eddy said.\u003c/p>\n\u003cp>The jury is set to begin deliberations on Monday. If they side with Musk, OpenAI and Microsoft could owe $150 billion in damages to be redirected to the nonprofit foundation, along with a court order dismantling OpenAI’s for-profit structure and removal of Altman and Brockman from their posts.\u003c/p>\n\u003cfigure id=\"attachment_12083616\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12083616\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne.jpg\" alt=\"\" width=\"2000\" height=\"1334\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/Gonzalez-Rogers-Yvonne-1536x1025.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Judge Yvonne Gonzalez Rogers of the U.S. District Court for the Northern District of California. \u003ccite>(Courtesy of Daily Journal)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>The jury will not have the final say, though. In a rare, but not unprecedented, move, U.S. District Judge Yvonne Gonzalez Rogers will have the ultimate right to rule on the claims.\u003c/p>\n\u003cp>According to Charlie Bullock, a senior research fellow at the Institute for Law and AI, this is because most times, “equitable claims” — breach of charitable trust and unjust enrichment — which involve non-monetary remedies, are decided by a judge.\u003c/p>\n\u003cp>In this case, Gonzalez Rogers elected to have an advisory jury, and Bullock said that typically, judges choose to go along with their decision.\u003c/p>\n\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12083612/lawyers-for-elon-musk-and-sam-altman-make-their-final-case-in-openai-trial",
"authors": [
"11913",
"251"
],
"categories": [
"news_31795",
"news_28250",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_1386",
"news_32668",
"news_3897",
"news_21891",
"news_34054",
"news_33542",
"news_33543",
"news_34586",
"news_1631"
],
"featImg": "news_12083803",
"label": "news"
},
"news_12083278": {
"type": "posts",
"id": "news_12083278",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12083278",
"score": null,
"sort": [
1778629278000
]
},
"guestAuthors": [],
"slug": "sam-altman-defends-himself-from-elon-musks-accusations-in-openai-trial",
"title": "Sam Altman Defends Himself From Elon Musk’s Accusations in OpenAI Trial",
"publishDate": 1778629278,
"format": "standard",
"headTitle": "Sam Altman Defends Himself From Elon Musk’s Accusations in OpenAI Trial | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>On the stand on Tuesday, OpenAI CEO Sam Altman said that Elon Musk tried to \u003ca href=\"https://www.kqed.org/news/12081916/are-elon-musk-and-openai-fighting-an-ai-arms-race-sam-altmans-lawyers-think-so\">wrest control over the company\u003c/a> they co-founded before the Tesla CEO’s 2018 exit.\u003c/p>\n\u003cp>Altman’s testimony in the federal trial in Oakland, which many see as a billionaire grudge match, pushed back on Musk’s claim that the powerful AI start-up betrayed its mission to benefit the public good. Musk has accused Altman of \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">“stealing a charity” \u003c/a>by building an $850 million for-profit company on the back of its nonprofit research lab.\u003c/p>\n\u003cp>Altman said that in early discussions about creating a for-profit arm, Musk sought majority ownership, and later proposed folding the nonprofit into his car company.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>“I read that as a lightweight threat,” Altman said of the plan to bring OpenAI into Tesla. “I don’t think it would have served the mission. I think it would have effectively destroyed the nonprofit in the process.”\u003c/p>\n\u003cp>“Mr. Musk did try to kill it, I guess twice,” he said.\u003c/p>\n\u003cp>As early as summer 2017, Altman, Musk and other OpenAI executives began discussing if and how to launch a for-profit, citing a need to raise more money to keep up with competitors like Google.\u003c/p>\n\u003cfigure id=\"attachment_12083394\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12083394 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman testifies in the trial in which Elon Musk claims that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 12, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Altman said they were “running the organization on a shoestring,” with a short runway of cash. To acquire the compute — or the GPUs and CPUs needed to power AI — and funding they needed to pursue artificial general intelligence, or a superintelligent AI technology known as AGI, the company would need more significant investments, the executives determined.\u003c/p>\n\u003cp>“I thought, of course, we needed to raise billions to quickly ramp,” he said. “I saw no way to do it.”\u003c/p>\n\u003cp>Altman, Greg Brockman, the president of OpenAI and Ilya Sutskever, a former top OpenAI computer scientist and member of its founding team, have said that in those conversations, Musk repeatedly proposed plans that would give him majority control. Initially, Altman said that he asked for 90% equity in a potential for-profit.\u003c/p>\n\u003cp>The other executives pushed back on this request, including in an email Altman sent to Musk at the time, in which he said, “I am worried about control. I don’t think any one person should have control of the world’s first AGI — in fact, the whole reason we started OpenAI is so that wouldn’t happen.”[aside postID=news_12083224 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty.jpg']Altman described Musk as “mercurial,” and said that when he left OpenAI in February 2018, after for-profit discussions fell apart, “people wondered if he’d try to take a vengeance on us” — which both he and his attorney, William Savitt, have alleged is exactly what Musk’s lawsuit aims to do.\u003c/p>\n\u003cp>During his cross-examination, though, Musk’s counsel Steven Molo seemed to suggest that it is Altman who has amassed significant control over OpenAI since it did launch a for-profit arm in 2019.\u003c/p>\n\u003cp>Molo asked Altman about the testimonies of various former OpenAI executives, who said he was untrustworthy and had a history of lying. Altman denied hearing those testimonies, but when asked if he had “repeatedly been called a liar” by people he has done business with, he said, “I have heard people say that.”\u003c/p>\n\u003cp>Molo said that Altman sits on the board of directors for both the OpenAI Foundation, the nonprofit arm, and OpenAI’s for-profit. He is also the company’s CEO.\u003c/p>\n\u003cp>“Would you ever fire yourself as the CEO of the for-profit?” Molo said, adding that the board of the nonprofit is supposed to provide oversight for the chief officer.\u003c/p>\n\u003cp>Altman said that CEOs are “almost always” on their company’s boards. When pressed, he said he had “no plans” to fire himself.\u003c/p>\n\u003cfigure id=\"attachment_12083294\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12083294\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Bret Taylor testifies in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 12, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Molo also asked Altman about how board members were selected following his brief firing in 2023. During the five-day ouster, there were long negotiations behind the scenes about whether Altman would return, and who would be on the board if he did. Altman, Brockman and other OpenAI executives who followed them out were also in discussions with Microsoft, OpenAI’s largest financial backer, which had offered to bring them on to start a new AI team.\u003c/p>\n\u003cp>Altman said initially he’d proposed to remove OpenAI’s board, which fired him, and replace it with four members, including himself. Altman was not made a board member at that time, but Molo said that he had proposed the three members who were ultimately selected — Bret Taylor, Larry Summers and Adam D’Angelo — in conversations with Microsoft CEO Satya Nadella.\u003c/p>\n\u003cp>Altman said that he had no power to appoint new board members, but that he did say which configurations he would be “willing” to be rehired into.\u003c/p>\n\u003cp>Earlier in the day, he characterized his return to OpenAI as running “back into a burning building to try to save it.”\u003c/p>\n\u003cp>Later this week, both Altman and Musk’s legal teams will present their closing arguments. Then the jury and judge will decide which tech leader to believe.\u003c/p>\n\u003cp>\u003c/p>\n",
"blocks": [],
"excerpt": "During a brief cross-examination of Altman, the Tesla CEO’s attorney questioned whether or not Altman was trustworthy.",
"status": "publish",
"parent": 0,
"modified": 1778630872,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 22,
"wordCount": 990
},
"headData": {
"title": "Sam Altman Defends Himself From Elon Musk’s Accusations in OpenAI Trial | KQED",
"description": "During a brief cross-examination of Altman, the Tesla CEO’s attorney questioned whether or not Altman was trustworthy.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Sam Altman Defends Himself From Elon Musk’s Accusations in OpenAI Trial",
"datePublished": "2026-05-12T16:41:18-07:00",
"dateModified": "2026-05-12T17:07:52-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"sticky": false,
"nprStoryId": "kqed-12083278",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12083278/sam-altman-defends-himself-from-elon-musks-accusations-in-openai-trial",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>On the stand on Tuesday, OpenAI CEO Sam Altman said that Elon Musk tried to \u003ca href=\"https://www.kqed.org/news/12081916/are-elon-musk-and-openai-fighting-an-ai-arms-race-sam-altmans-lawyers-think-so\">wrest control over the company\u003c/a> they co-founded before the Tesla CEO’s 2018 exit.\u003c/p>\n\u003cp>Altman’s testimony in the federal trial in Oakland, which many see as a billionaire grudge match, pushed back on Musk’s claim that the powerful AI start-up betrayed its mission to benefit the public good. Musk has accused Altman of \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">“stealing a charity” \u003c/a>by building an $850 million for-profit company on the back of its nonprofit research lab.\u003c/p>\n\u003cp>Altman said that in early discussions about creating a for-profit arm, Musk sought majority ownership, and later proposed folding the nonprofit into his car company.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>“I read that as a lightweight threat,” Altman said of the plan to bring OpenAI into Tesla. “I don’t think it would have served the mission. I think it would have effectively destroyed the nonprofit in the process.”\u003c/p>\n\u003cp>“Mr. Musk did try to kill it, I guess twice,” he said.\u003c/p>\n\u003cp>As early as summer 2017, Altman, Musk and other OpenAI executives began discussing if and how to launch a for-profit, citing a need to raise more money to keep up with competitors like Google.\u003c/p>\n\u003cfigure id=\"attachment_12083394\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12083394 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman testifies in the trial in which Elon Musk claims that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 12, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Altman said they were “running the organization on a shoestring,” with a short runway of cash. To acquire the compute — or the GPUs and CPUs needed to power AI — and funding they needed to pursue artificial general intelligence, or a superintelligent AI technology known as AGI, the company would need more significant investments, the executives determined.\u003c/p>\n\u003cp>“I thought, of course, we needed to raise billions to quickly ramp,” he said. “I saw no way to do it.”\u003c/p>\n\u003cp>Altman, Greg Brockman, the president of OpenAI and Ilya Sutskever, a former top OpenAI computer scientist and member of its founding team, have said that in those conversations, Musk repeatedly proposed plans that would give him majority control. Initially, Altman said that he asked for 90% equity in a potential for-profit.\u003c/p>\n\u003cp>The other executives pushed back on this request, including in an email Altman sent to Musk at the time, in which he said, “I am worried about control. I don’t think any one person should have control of the world’s first AGI — in fact, the whole reason we started OpenAI is so that wouldn’t happen.”\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12083224",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/SamAltmanGetty.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Altman described Musk as “mercurial,” and said that when he left OpenAI in February 2018, after for-profit discussions fell apart, “people wondered if he’d try to take a vengeance on us” — which both he and his attorney, William Savitt, have alleged is exactly what Musk’s lawsuit aims to do.\u003c/p>\n\u003cp>During his cross-examination, though, Musk’s counsel Steven Molo seemed to suggest that it is Altman who has amassed significant control over OpenAI since it did launch a for-profit arm in 2019.\u003c/p>\n\u003cp>Molo asked Altman about the testimonies of various former OpenAI executives, who said he was untrustworthy and had a history of lying. Altman denied hearing those testimonies, but when asked if he had “repeatedly been called a liar” by people he has done business with, he said, “I have heard people say that.”\u003c/p>\n\u003cp>Molo said that Altman sits on the board of directors for both the OpenAI Foundation, the nonprofit arm, and OpenAI’s for-profit. He is also the company’s CEO.\u003c/p>\n\u003cp>“Would you ever fire yourself as the CEO of the for-profit?” Molo said, adding that the board of the nonprofit is supposed to provide oversight for the chief officer.\u003c/p>\n\u003cp>Altman said that CEOs are “almost always” on their company’s boards. When pressed, he said he had “no plans” to fire himself.\u003c/p>\n\u003cfigure id=\"attachment_12083294\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12083294\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260512-MUSK-ALTMAN-TRIAL-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Bret Taylor testifies in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity rather than solely for profit in Oakland on May 12, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Molo also asked Altman about how board members were selected following his brief firing in 2023. During the five-day ouster, there were long negotiations behind the scenes about whether Altman would return, and who would be on the board if he did. Altman, Brockman and other OpenAI executives who followed them out were also in discussions with Microsoft, OpenAI’s largest financial backer, which had offered to bring them on to start a new AI team.\u003c/p>\n\u003cp>Altman said initially he’d proposed to remove OpenAI’s board, which fired him, and replace it with four members, including himself. Altman was not made a board member at that time, but Molo said that he had proposed the three members who were ultimately selected — Bret Taylor, Larry Summers and Adam D’Angelo — in conversations with Microsoft CEO Satya Nadella.\u003c/p>\n\u003cp>Altman said that he had no power to appoint new board members, but that he did say which configurations he would be “willing” to be rehired into.\u003c/p>\n\u003cp>Earlier in the day, he characterized his return to OpenAI as running “back into a burning building to try to save it.”\u003c/p>\n\u003cp>Later this week, both Altman and Musk’s legal teams will present their closing arguments. Then the jury and judge will decide which tech leader to believe.\u003c/p>\n\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12083278/sam-altman-defends-himself-from-elon-musks-accusations-in-openai-trial",
"authors": [
"11913",
"251"
],
"categories": [
"news_6188",
"news_28250",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_1386",
"news_32668",
"news_3897",
"news_27626",
"news_21891",
"news_34054",
"news_33542",
"news_33543",
"news_34586",
"news_1631"
],
"featImg": "news_12083392",
"label": "news"
},
"news_12083224": {
"type": "posts",
"id": "news_12083224",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12083224",
"score": null,
"sort": [
1778546112000
]
},
"guestAuthors": [],
"slug": "former-openai-exec-calls-decision-to-remove-sam-altman-a-hail-mary-during-musk-trial",
"title": "Former OpenAI Exec Calls Decision to Remove Sam Altman a ‘Hail Mary’ During Musk Trial",
"publishDate": 1778546112,
"format": "standard",
"headTitle": "Former OpenAI Exec Calls Decision to Remove Sam Altman a ‘Hail Mary’ During Musk Trial | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>Microsoft’s CEO and another major player took the stand on Monday in \u003ca href=\"https://www.kqed.org/news/tag/oakland\">Oakland\u003c/a>, testifying in the blockbuster trial between OpenAI co-founders Elon Musk and Sam Altman.\u003c/p>\n\u003cp>Ahead of Altman’s testimony, Musk’s attorney Steven Molo questioned Microsoft CEO Satya Nadella and Ilya Sutskever, a top OpenAI computer scientist who departed the company in 2024. Sutskever discussed his role in orchestrating Altman’s brief ouster in 2023.\u003c/p>\n\u003cp>Over five days in November 2023, Altman was removed and reinstated from his post, after a coalition of board members raised concerns that he had not been “consistently candid in his communications” and cited a breakdown of trust.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>Whether Altman and other executives have maintained OpenAI’s initial stated mission — to develop AI safely and for the “benefit of humanity” — is critical to Musk’s suit, which claims that leaders breached their duty to its nonprofit mission by building a for-profit company on top of it. Musk also alleged that the company unfairly benefited at his expense.\u003c/p>\n\u003cp>Musk also alleges that Microsoft, which is OpenAI’s largest financial backer and until this week held the exclusive rights to license and sell its technology, aided and abetted that breach of trust.\u003c/p>\n\u003cp>Molo questioned Nadella about Microsoft’s motive to invest in OpenAI — a $13 billion input that Nadella said is expected to see a return of about $92 billion, “if it works out.”\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12081686 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland, on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Musk’s attorney pointed out Nadella’s fiduciary duty to maximize profit, and referenced a series of texts between him and Altman that appeared to show Nadella pushing for an earlier rollout of the paid version of ChatGPT.\u003c/p>\n\u003cp>“When chatGPT paid?” Nadella wrote in the message.\u003c/p>\n\u003cp>Altman said that there was “Not enough compute to make it a good consumer experience,” to which Nadella said, “The sooner the better.”\u003c/p>\n\u003cp>Nadella said that the reason Microsoft invested was that OpenAI was pursuing a for-profit model, but he said, “If the pie became larger, the nonprofit would benefit as well.”[aside postID=news_12081916 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/AP26118555622828-2000x1333.jpg']Molo asked Nadella if he was aware that, for a period of time, OpenAI’s nonprofit did not have any employees.\u003c/p>\n\u003cp>“I am not,” Nadella said.\u003c/p>\n\u003cp>Molo also questioned Nadella about Microsoft’s role during Altman’s brief ouster. At the time, Nadella announced that he would hire Altman, along with OpenAI’s third co-founder and current president, Greg Brockman, as well as other allies, to head up a new AI team at Microsoft.\u003c/p>\n\u003cp>Nadella said that he “had ideas about how Sam [Altman] and the other employees could join Microsoft if they were not reinstated.”\u003c/p>\n\u003cp>“If people were going to leave OpenAI, I wanted them to come to Microsoft,” he said.\u003c/p>\n\u003cp>Molo asked Nadella if he knew why Altman had been removed, to which Nadella said he was never given an “explicit answer.”\u003c/p>\n\u003cp>“Did the thought occur to you … the board might issue a public statement about why they fired Altman?” Molo said.\u003c/p>\n\u003cp>Nadella said during that period — referred to as “The Blip” by many OpenAI employees — he was focused on ensuring continuity for customers.\u003c/p>\n\u003cp>“It goes back to me wanting to communicate to customers that they can count on us,” he said. “Come Monday, that doesn’t just disappear.”\u003c/p>\n\u003cfigure id=\"attachment_12082325\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082325 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman watches as OpenAI President Greg Brockman testifies in the trial in which Elon Musk claims that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland, on May 4, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Sutskever, who took the stand after Nadella, described Altman’s removal differently. He said it was a “Hail Mary” to save OpenAI, which had become an environment that was “not conducive” to the technology’s safety.\u003c/p>\n\u003cp>“I felt a great deal of ownership of OpenAI,” he said. “I felt like I created this company. I simply cared for it, and I didn’t want it to be destroyed.”\u003c/p>\n\u003cp>Sutskever, who helped lead the ouster, had compiled a more than 50-page record of Altman’s “consistent pattern of lying,” including misrepresenting facts, safety protocols and company information to the board and executives.\u003c/p>\n\u003cp>Sutskever maintained that he had worked on a team that aimed to focus on long-term risks as more powerful AI was built.\u003c/p>\n\u003cp>“The goal of the super alignment is to do the research in advance, such that humanity will have the technological means to make it controlled and safe,” he said.\u003c/p>\n\u003cp>The team was disbanded days after he departed the company, in May 2024.\u003c/p>\n\u003cp>\u003c/p>\n",
"blocks": [],
"excerpt": "The testimonies on Monday centered on Sam Altman’s brief 2023 ousting from OpenAI, as allegations mounted against the tech giant’s conduct and Microsoft’s motives in backing the AI company.",
"status": "publish",
"parent": 0,
"modified": 1778547375,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 26,
"wordCount": 890
},
"headData": {
"title": "Former OpenAI Exec Calls Decision to Remove Sam Altman a ‘Hail Mary’ During Musk Trial | KQED",
"description": "The testimonies on Monday centered on Sam Altman’s brief 2023 ousting from OpenAI, as allegations mounted against the tech giant’s conduct and Microsoft’s motives in backing the AI company.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Former OpenAI Exec Calls Decision to Remove Sam Altman a ‘Hail Mary’ During Musk Trial",
"datePublished": "2026-05-11T17:35:12-07:00",
"dateModified": "2026-05-11T17:56:15-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"sticky": false,
"nprStoryId": "kqed-12083224",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12083224/former-openai-exec-calls-decision-to-remove-sam-altman-a-hail-mary-during-musk-trial",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>Microsoft’s CEO and another major player took the stand on Monday in \u003ca href=\"https://www.kqed.org/news/tag/oakland\">Oakland\u003c/a>, testifying in the blockbuster trial between OpenAI co-founders Elon Musk and Sam Altman.\u003c/p>\n\u003cp>Ahead of Altman’s testimony, Musk’s attorney Steven Molo questioned Microsoft CEO Satya Nadella and Ilya Sutskever, a top OpenAI computer scientist who departed the company in 2024. Sutskever discussed his role in orchestrating Altman’s brief ouster in 2023.\u003c/p>\n\u003cp>Over five days in November 2023, Altman was removed and reinstated from his post, after a coalition of board members raised concerns that he had not been “consistently candid in his communications” and cited a breakdown of trust.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>Whether Altman and other executives have maintained OpenAI’s initial stated mission — to develop AI safely and for the “benefit of humanity” — is critical to Musk’s suit, which claims that leaders breached their duty to its nonprofit mission by building a for-profit company on top of it. Musk also alleged that the company unfairly benefited at his expense.\u003c/p>\n\u003cp>Musk also alleges that Microsoft, which is OpenAI’s largest financial backer and until this week held the exclusive rights to license and sell its technology, aided and abetted that breach of trust.\u003c/p>\n\u003cp>Molo questioned Nadella about Microsoft’s motive to invest in OpenAI — a $13 billion input that Nadella said is expected to see a return of about $92 billion, “if it works out.”\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12081686 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland, on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Musk’s attorney pointed out Nadella’s fiduciary duty to maximize profit, and referenced a series of texts between him and Altman that appeared to show Nadella pushing for an earlier rollout of the paid version of ChatGPT.\u003c/p>\n\u003cp>“When chatGPT paid?” Nadella wrote in the message.\u003c/p>\n\u003cp>Altman said that there was “Not enough compute to make it a good consumer experience,” to which Nadella said, “The sooner the better.”\u003c/p>\n\u003cp>Nadella said that the reason Microsoft invested was that OpenAI was pursuing a for-profit model, but he said, “If the pie became larger, the nonprofit would benefit as well.”\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12081916",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/AP26118555622828-2000x1333.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Molo asked Nadella if he was aware that, for a period of time, OpenAI’s nonprofit did not have any employees.\u003c/p>\n\u003cp>“I am not,” Nadella said.\u003c/p>\n\u003cp>Molo also questioned Nadella about Microsoft’s role during Altman’s brief ouster. At the time, Nadella announced that he would hire Altman, along with OpenAI’s third co-founder and current president, Greg Brockman, as well as other allies, to head up a new AI team at Microsoft.\u003c/p>\n\u003cp>Nadella said that he “had ideas about how Sam [Altman] and the other employees could join Microsoft if they were not reinstated.”\u003c/p>\n\u003cp>“If people were going to leave OpenAI, I wanted them to come to Microsoft,” he said.\u003c/p>\n\u003cp>Molo asked Nadella if he knew why Altman had been removed, to which Nadella said he was never given an “explicit answer.”\u003c/p>\n\u003cp>“Did the thought occur to you … the board might issue a public statement about why they fired Altman?” Molo said.\u003c/p>\n\u003cp>Nadella said during that period — referred to as “The Blip” by many OpenAI employees — he was focused on ensuring continuity for customers.\u003c/p>\n\u003cp>“It goes back to me wanting to communicate to customers that they can count on us,” he said. “Come Monday, that doesn’t just disappear.”\u003c/p>\n\u003cfigure id=\"attachment_12082325\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082325 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/260504-MUSK-ALTMAN-VB-03-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman watches as OpenAI President Greg Brockman testifies in the trial in which Elon Musk claims that Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland, on May 4, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Sutskever, who took the stand after Nadella, described Altman’s removal differently. He said it was a “Hail Mary” to save OpenAI, which had become an environment that was “not conducive” to the technology’s safety.\u003c/p>\n\u003cp>“I felt a great deal of ownership of OpenAI,” he said. “I felt like I created this company. I simply cared for it, and I didn’t want it to be destroyed.”\u003c/p>\n\u003cp>Sutskever, who helped lead the ouster, had compiled a more than 50-page record of Altman’s “consistent pattern of lying,” including misrepresenting facts, safety protocols and company information to the board and executives.\u003c/p>\n\u003cp>Sutskever maintained that he had worked on a team that aimed to focus on long-term risks as more powerful AI was built.\u003c/p>\n\u003cp>“The goal of the super alignment is to do the research in advance, such that humanity will have the technological means to make it controlled and safe,” he said.\u003c/p>\n\u003cp>The team was disbanded days after he departed the company, in May 2024.\u003c/p>\n\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12083224/former-openai-exec-calls-decision-to-remove-sam-altman-a-hail-mary-during-musk-trial",
"authors": [
"11913",
"251"
],
"categories": [
"news_31795",
"news_6188",
"news_28250",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_1386",
"news_32668",
"news_3897",
"news_27626",
"news_19954",
"news_21891",
"news_34054",
"news_33542",
"news_33543",
"news_34586",
"news_1631"
],
"featImg": "news_12083235",
"label": "news"
},
"news_12082428": {
"type": "posts",
"id": "news_12082428",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12082428",
"score": null,
"sort": [
1778061636000
]
},
"guestAuthors": [],
"slug": "inside-sam-altman-and-elon-musks-battle-over-openai",
"title": "Inside Elon Musk and Sam Altman's Battle Over OpenAI",
"publishDate": 1778061636,
"format": "audio",
"headTitle": "Inside Elon Musk and Sam Altman’s Battle Over OpenAI | KQED",
"labelTerm": {},
"content": "\u003cp>\u003cspan style=\"font-weight: 400\">Jurors and journalists are getting a peek into the world of OpenAI and its founding as two of the richest, most powerful men in tech duke it out in an Oakland federal courthouse. \u003c/span>\u003c/p>\n\u003cp>\u003cspan style=\"font-weight: 400\">Elon Musk claims that Sam Altman and other co-founders of OpenAI abandoned their founding promise to develop AI for the benefit of humanity. But does anyone here really have our best interests at heart? KQED’s Rachael Myrow takes us inside.\u003c/span>\u003c/p>\n\u003cp>\u003cstrong>Links:\u003c/strong>\u003c/p>\n\u003cul>\n\u003cli style=\"list-style-type: none\">\n\u003cul>\n\u003cli>\u003ca href=\"https://www.kqed.org/news/12081290/how-to-unscramble-an-omelet-in-silicon-valley-the-musk-v-altman-trial-that-will-try\">How to Unscramble an Omelet in Silicon Valley: The Musk v. Altman Trial That Will Try\u003c/a>\u003c/li>\n\u003c/ul>\n\u003c/li>\n\u003c/ul>\n\u003cp>\u003c!-- iframe plugin v.4.3 wordpress.org/plugins/iframe/ -->\u003cbr>\n\u003ciframe loading=\"lazy\" frameborder=\"0\" height=\"200\" scrolling=\"no\" src=\"https://playlist.megaphone.fm?e=KQINC4004396119\" width=\"100%\" class=\"iframe-class\">\u003c/iframe>\u003c/p>\n\u003ch3>\u003cstrong>Episode Transcript\u003c/strong>\u003c/h3>\n\u003cp>\u003cem>This is a computer-generated transcript. While our team has reviewed it, there may be errors.\u003c/em>\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:00:49] \u003c/em>I’m Ericka Cruz-Gavarra and welcome to The Bay, local news to keep you rooted. Inside a federal courthouse in downtown Oakland, in front of a judge and a jury of their peers, two of the most powerful men in the world are duking it out in court over whether OpenAI, the company behind ChatGPT, was built on a lie. Elon Musk is suing OpenAI and its CEO, Sam Altman. For abandoning their founding promise to develop AI for the benefit of humanity. And whether or not you actually believe any of them really had our best interests in mind, one thing is true, that the battle over who runs AI is all about ego and power.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:01:47] \u003c/em>No matter which side wins, the people are going to lose because they are not doing this actually for the benefit of humanity, it’s not about ethics, this is all about power plays within an unfettered, unregulated AI scheme.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:01] \u003c/em>Today, KQED’s Rachael Myrow takes us inside the OpenAI trial.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:19] \u003c/em>It seems like you’re in a pretty dynamic scene right now, Rachael. Can you actually tell us where you are?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:02:26] \u003c/em>I’m outside the federal courthouse in Oakland where Musk v. Altman et al. Is playing out.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:36] \u003c/em>Rachael Myrow is a senior editor at KQED.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:02:41] \u003c/em>This plaza is right on the street so you hear the chirping every time somebody presses a button to cross the street. You hear garbage trucks rolling past. Inside the courtroom is presided over by Judge Yvonne Gonzalez Rogers and it is packed every single day. Armies of lawyers of course but also journalists from across the country, even a couple from France. And some members of the public. I’d like to call this the hottest theater ticket in Silicon Valley. We got to see Elon Musk spend four days on the witness stand. Sam Altman is sitting just a few feet away in the defense section. These two men genuinely cannot stand each other.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:03:30] \u003c/em>And I understand, Rachel, that there’s not just folks inside of the courtroom for this trial, but also outside protesting as well.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:03:41] \u003c/em>Yes, on the very first day, actually when jury selection was taking place, protesters gathered in large numbers outside the courthouse on the plaza with some very pointed and colorful signs.\u003c/p>\n\u003cp>\u003cb>Valerie Sizemore: \u003c/b>\u003cem>[00:03:57] \u003c/em>I used to be a software engineer, but have been unemployed by AI. So now I’m trying to make the resistance happen.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:06] \u003c/em>I talked to one protester, Valerie Sizemore of Berkeley, who kind of represented, I think, a lot of Bayarians.\u003c/p>\n\u003cp>\u003cb>Valerie Sizemore: \u003c/b>\u003cem>[00:04:15] \u003c/em>I’m not here because I care about the outcome of this trial. I really don’t care. I hope it’s really expensive for someone and like hurts both companies as much as possible.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:04:27] \u003c/em>Yeah, and it’s, I guess, two-for-one for her to just be outside the courthouse protesting the both of them.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:34] \u003c/em>Exactly.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:04:39] \u003c/em>Well, Rachael, I wanna step back a little bit and talk about this trial and just how we even got here. I mean, remind us who is on trial and what exactly these two are fighting over?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:53] \u003c/em>So it’s a little more than two people. Elon Musk is suing Sam Altman and also Greg Brockman, who is OpenAI’s co-founder and president. Musk is sueing OpenAI itself and also Microsoft, which invested $13 billion in OpenAI after Musk left.\u003c/p>\n\u003cp>\u003cb>Interviewer: \u003c/b>\u003cem>[00:05:18] \u003c/em>All right, we’re gonna wrap up the day. I’m gonna do a fireside chat with Sam Altman.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:05:25] \u003c/em>Let’s dial the clock all the way back to 2015. Musk and Altman found OpenAI as a nonprofit explicitly to develop artificial general intelligence safely and for the benefit of all humanity.\u003c/p>\n\u003cp>\u003cb>Sam Altman: \u003c/b>\u003cem>[00:05:44] \u003c/em>You know, I think AI will probably, like most likely, sort of lead to the end of the world, but in the meantime, there will be great companies created with serious machine learning. I actually just agreed to fund a company that is not even really a company, sort of a semi-company, semi-nonprofit, doing AI safety research.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:06:04] \u003c/em>At some point shortly thereafter, it became clear to all parties involved, including Musk, that they needed to establish a for-profit arm as well in order to raise money to pay for things like computing power for this very energy-intensive computer software and also to bring in talent, to bring the best minds of the industry. Musk’s lawsuit is arguing that thereabouts Altman and other co-founders of OpenAI, because there were other people involved, betrayed the mission, that they were actually in it for the profit.\u003c/p>\n\u003cp>\u003cb>Interviewer: \u003c/b>\u003cem>[00:06:45] \u003c/em>Open AI, I mean you seem somewhat frustrated with them. You were one of the big contributors early on?\u003c/p>\n\u003cp>\u003cb>Elon Musk: \u003c/b>\u003cem>[00:06:49] \u003c/em>The reason, I am the reason Open AI exists.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:06:54] \u003c/em>So he wants more than his money back. He wants Altman and OpenAI’s co-founder and president, Greg Brockman, taken off the board. And he wants $130 billion, disgorged by the for-profit and handed over to the non-profit. The word charity, Ericka, doesn’t appear once in OpenAI’s founding blog post, but Musk keeps referring to OpenAI as a charity. But as OpenAI lawyers like to point out, Musk left OpenAI and then he launched his own AI venture, XAI, which is not a nonprofit and arguably does not operate for the benefit of humanity, for which it has been sued repeatedly.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:07:50] \u003c/em>So it sounds like Elon Musk is basically saying they stole his charity, and Sam Altman is saying, ‘You chose to walk away.’\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:08:02] \u003c/em>Yeah. That’s it in a nutshell. There was this funny moment when Musk was on the witness stand. He looked at the jury and he said, quote, it’s not OK to steal a charity. And then he predicted that if Open AI wins this case, the face of charity law in America could be altered forever. At some point, the judge broke in and said, let’s remind the jury, you’re not a lawyer. She’s talking to Musk. And then he replied. I did take Law 101, which got a laugh out of most people in the court.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:08:36] \u003c/em>Geez.\u003c/p>\n\u003cp>\u003cb>\u003c/b>\u003cem>[00:08:39] \u003c/em>Rachael, what do we make of Sam Altman’s role in this? It sounds like Elon Musk is saying that Sam Altman lied to him.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:08:46] \u003c/em>That is a very good question. I need to mention here that we have not seen Sam Altman take the stand in this trial yet. So Altman has not yet had the chance to make his case. Just a few weeks ago, we saw a comprehensive profile of Sam Altman in the New Yorker magazine talking to lots and lots of people that Sam Altmann is an inveterate liar, the kind of person who will tell you what you want to hear and then go back on it. We haven’t had the opportunity yet to really get into what his character was like during the early days of OpenAI, but pretty much everyone in that courtroom has read that article.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:09:47] \u003c/em>Coming up, what the OpenAI trial is really about. Stay with us.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:10:38] \u003c/em>I mean, Rachael, I gotta say, as I’m reading these stories about this case, it really just sounds like a fight between two of some of the richest billionaires in Silicon Valley over this company that they co-founded. But obviously, what’s at the center of it and what is at stake is this very powerful technology that even they seem to acknowledge has the potential to change the world. So what do you think this is really about?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:11:13] \u003c/em>Clearly about power, clearly about money, clearly about market dominance. And I do want to say that even though the judge is saying we are not going to talk about the AI apocalypse, it is something that is genuinely on the minds of all of these people in the industry in Silicon Valley and also the rest of us, right? I mean there are people here who take AI safety seriously. Who also think OpenAI has drifted dangerously from its mission. I mean, we’ve seen bad actors using the software who have upended the labor market, terrified all of us from a cybersecurity perspective, made it impossible to get redress as a customer and sometimes as a citizen, enabled a surveillance state here and abroad. I mean I could go on, Ericka, because It’s 100% clear to us and the people building this software that there’s a race to the bottom going on from a moral perspective.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:12:30] \u003c/em>I do want to ask you this question, Rachel, because Elon Musk is saying in this trial that he is the one standing up for the public on AI. Rachel, is there someone working in the public interest when it comes to AI and holding AI companies accountable?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:12:53] \u003c/em>Well, you know, don’t we wish? He’d like to present himself as thinking pro-human first, but you know, he also created XAI and has reportedly personally directed his engineers to make XAI a manifestly unsafe product. The judge noted the irony out loud. She said to Musk’s attorneys at one point, It is ironic that your client, despite these risks, is creating a company in the exact same space. And then she added, and I just thought this was so remarkable, coming from, again, a sitting federal judge, quote, I suspect there are people who don’t want to put the future in Mr. Musk’s hands, unquote.\u003c/p>\n\u003cp>\u003cb>Jill Horowitz: \u003c/b>\u003cem>[00:13:44] \u003c/em>And in that sense, I don’t understand why Musk is the one who gets to ask that question. Jill Horowitz, who specifically specializes in non-profit law. At Northwestern’s law school put it this way when parties have this much money and this much power they can trample over conventional protections of the public interest\u003c/p>\n\u003cp>\u003cb>Jill Horowitz: \u003c/b>\u003cem>[00:14:07] \u003c/em>We’ve got a CEO who is a very powerful player. And then we have this outside party who’s purported to be thinking about the best interest of the nonprofit, but he’s a competitor.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:14:22] \u003c/em>Congress, you don’t need me to tell you, hasn’t passed any meaningful federal AI regulation. The Trump administration is lobbying alongside the lobbyists for unfettered freedom for the AI industry. And so we end up here, Ericka, in a federal courthouse in Oakland watching two billionaires fight over their recent past. This trial gives us a window into the wheeling and dealing. But it doesn’t give us any power to change the trajectory of AI.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:14:54] \u003c/em>Rachael what happens if if either Elon Musk or Sam Altman wins this trial?\u003c/p>\n\u003cp>\u003cb>\u003c/b>\u003cem>[00:15:01] \u003c/em>So if Musk wins, Judge Gonzalez-Rogers could order OpenAI to revert to a non-profit structure, remove Altman and Brockman, direct some $130 billion in gains back to the non- profit foundation. That would be legally unprecedented and would certainly send shockwaves throughout Silicon Valley. If OpenAI wins, the restructuring stands, the IPO proceeds. And the message to the industry is essentially, you can do this too. You can take a non-profit, make it nominally in charge of a for-profit arm that you build into a trillion-dollar company, and the legal system won’t stop you.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:15:52] \u003c/em>Last question for you, Rachael. For the protesters outside, what do you think they want to see happen? And do you they care here about who wins?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:16:03] \u003c/em>My sense from talking with Ashley Ortiz, who was one of the organizers of the first and biggest protest outside, is that for a lot of the people out here carrying signs and wearing t-shirts that say stop AI, neither Musk nor Altman represents their interests and by extension the public’s interests.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:16:28] \u003c/em>Decision everyone sucks here and y’all both need to take responsibility for your part in this crappy situation.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:16:34] \u003c/em>They want accountability for AI, period. I don’t know if they actually think they’re gonna get what they’re asking for, but they wanna make a noise while they can.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:16:46] \u003c/em>We’re letting them both know that both sides, no matter which side wins, the people are going to lose because they are not doing this actually for the benefit of humanity. It’s not about ethics. This is all about power plays within an unfettered, unregulated AI scape.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:17:06] \u003c/em>These are the models that are changing our world, and they’re doing it now. And regardless of whether OpenAI survives this trial, we’re still gonna have the world that OpenAI helped to create.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:17:28] \u003c/em>Well, Rachael, thank you so much for chatting with me outside the courtroom and for making the time in your busy morning, I appreciate it.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:17:36] \u003c/em>You bet.\u003c/p>\n\u003cp>[ad floatright]\u003c/p>\n\u003cp>\u003ci>\u003cspan style=\"font-weight: 400\">Some members of the KQED podcast team are represented by The Screen Actors Guild, American Federation of Television and Radio Artists, San Francisco-Northern California Local.\u003c/span>\u003c/i>\u003c/p>\n\n",
"blocks": [],
"excerpt": "At issue is whether Sam Altman abandoned his founding promise with Elon Musk to develop AI for the benefit of humanity.",
"status": "publish",
"parent": 0,
"modified": 1778097044,
"stats": {
"hasAudio": true,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 54,
"wordCount": 2552
},
"headData": {
"title": "Inside Elon Musk and Sam Altman's Battle Over OpenAI | KQED",
"description": "At issue is whether Sam Altman abandoned his founding promise with Elon Musk to develop AI for the benefit of humanity.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Inside Elon Musk and Sam Altman's Battle Over OpenAI",
"datePublished": "2026-05-06T03:00:36-07:00",
"dateModified": "2026-05-06T12:50:44-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 8,
"slug": "news",
"name": "News"
},
"source": "The Bay",
"sourceUrl": "https://www.kqed.org/podcasts/thebay",
"audioUrl": "https://www.podtrac.com/pts/redirect.mp3/traffic.megaphone.fm/KQINC4004396119.mp3",
"sticky": false,
"nprStoryId": "kqed-12082428",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12082428/inside-sam-altman-and-elon-musks-battle-over-openai",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>\u003cspan style=\"font-weight: 400\">Jurors and journalists are getting a peek into the world of OpenAI and its founding as two of the richest, most powerful men in tech duke it out in an Oakland federal courthouse. \u003c/span>\u003c/p>\n\u003cp>\u003cspan style=\"font-weight: 400\">Elon Musk claims that Sam Altman and other co-founders of OpenAI abandoned their founding promise to develop AI for the benefit of humanity. But does anyone here really have our best interests at heart? KQED’s Rachael Myrow takes us inside.\u003c/span>\u003c/p>\n\u003cp>\u003cstrong>Links:\u003c/strong>\u003c/p>\n\u003cul>\n\u003cli style=\"list-style-type: none\">\n\u003cul>\n\u003cli>\u003ca href=\"https://www.kqed.org/news/12081290/how-to-unscramble-an-omelet-in-silicon-valley-the-musk-v-altman-trial-that-will-try\">How to Unscramble an Omelet in Silicon Valley: The Musk v. Altman Trial That Will Try\u003c/a>\u003c/li>\n\u003c/ul>\n\u003c/li>\n\u003c/ul>\n\u003cp>\u003c!-- iframe plugin v.4.3 wordpress.org/plugins/iframe/ -->\u003cbr>\n\u003ciframe loading=\"lazy\" frameborder=\"0\" height=\"200\" scrolling=\"no\" src=\"https://playlist.megaphone.fm?e=KQINC4004396119\" width=\"100%\" class=\"iframe-class\">\u003c/iframe>\u003c/p>\n\u003ch3>\u003cstrong>Episode Transcript\u003c/strong>\u003c/h3>\n\u003cp>\u003cem>This is a computer-generated transcript. While our team has reviewed it, there may be errors.\u003c/em>\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:00:49] \u003c/em>I’m Ericka Cruz-Gavarra and welcome to The Bay, local news to keep you rooted. Inside a federal courthouse in downtown Oakland, in front of a judge and a jury of their peers, two of the most powerful men in the world are duking it out in court over whether OpenAI, the company behind ChatGPT, was built on a lie. Elon Musk is suing OpenAI and its CEO, Sam Altman. For abandoning their founding promise to develop AI for the benefit of humanity. And whether or not you actually believe any of them really had our best interests in mind, one thing is true, that the battle over who runs AI is all about ego and power.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:01:47] \u003c/em>No matter which side wins, the people are going to lose because they are not doing this actually for the benefit of humanity, it’s not about ethics, this is all about power plays within an unfettered, unregulated AI scheme.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:01] \u003c/em>Today, KQED’s Rachael Myrow takes us inside the OpenAI trial.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:19] \u003c/em>It seems like you’re in a pretty dynamic scene right now, Rachael. Can you actually tell us where you are?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:02:26] \u003c/em>I’m outside the federal courthouse in Oakland where Musk v. Altman et al. Is playing out.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:02:36] \u003c/em>Rachael Myrow is a senior editor at KQED.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:02:41] \u003c/em>This plaza is right on the street so you hear the chirping every time somebody presses a button to cross the street. You hear garbage trucks rolling past. Inside the courtroom is presided over by Judge Yvonne Gonzalez Rogers and it is packed every single day. Armies of lawyers of course but also journalists from across the country, even a couple from France. And some members of the public. I’d like to call this the hottest theater ticket in Silicon Valley. We got to see Elon Musk spend four days on the witness stand. Sam Altman is sitting just a few feet away in the defense section. These two men genuinely cannot stand each other.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:03:30] \u003c/em>And I understand, Rachel, that there’s not just folks inside of the courtroom for this trial, but also outside protesting as well.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:03:41] \u003c/em>Yes, on the very first day, actually when jury selection was taking place, protesters gathered in large numbers outside the courthouse on the plaza with some very pointed and colorful signs.\u003c/p>\n\u003cp>\u003cb>Valerie Sizemore: \u003c/b>\u003cem>[00:03:57] \u003c/em>I used to be a software engineer, but have been unemployed by AI. So now I’m trying to make the resistance happen.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:06] \u003c/em>I talked to one protester, Valerie Sizemore of Berkeley, who kind of represented, I think, a lot of Bayarians.\u003c/p>\n\u003cp>\u003cb>Valerie Sizemore: \u003c/b>\u003cem>[00:04:15] \u003c/em>I’m not here because I care about the outcome of this trial. I really don’t care. I hope it’s really expensive for someone and like hurts both companies as much as possible.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:04:27] \u003c/em>Yeah, and it’s, I guess, two-for-one for her to just be outside the courthouse protesting the both of them.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:34] \u003c/em>Exactly.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:04:39] \u003c/em>Well, Rachael, I wanna step back a little bit and talk about this trial and just how we even got here. I mean, remind us who is on trial and what exactly these two are fighting over?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:04:53] \u003c/em>So it’s a little more than two people. Elon Musk is suing Sam Altman and also Greg Brockman, who is OpenAI’s co-founder and president. Musk is sueing OpenAI itself and also Microsoft, which invested $13 billion in OpenAI after Musk left.\u003c/p>\n\u003cp>\u003cb>Interviewer: \u003c/b>\u003cem>[00:05:18] \u003c/em>All right, we’re gonna wrap up the day. I’m gonna do a fireside chat with Sam Altman.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:05:25] \u003c/em>Let’s dial the clock all the way back to 2015. Musk and Altman found OpenAI as a nonprofit explicitly to develop artificial general intelligence safely and for the benefit of all humanity.\u003c/p>\n\u003cp>\u003cb>Sam Altman: \u003c/b>\u003cem>[00:05:44] \u003c/em>You know, I think AI will probably, like most likely, sort of lead to the end of the world, but in the meantime, there will be great companies created with serious machine learning. I actually just agreed to fund a company that is not even really a company, sort of a semi-company, semi-nonprofit, doing AI safety research.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:06:04] \u003c/em>At some point shortly thereafter, it became clear to all parties involved, including Musk, that they needed to establish a for-profit arm as well in order to raise money to pay for things like computing power for this very energy-intensive computer software and also to bring in talent, to bring the best minds of the industry. Musk’s lawsuit is arguing that thereabouts Altman and other co-founders of OpenAI, because there were other people involved, betrayed the mission, that they were actually in it for the profit.\u003c/p>\n\u003cp>\u003cb>Interviewer: \u003c/b>\u003cem>[00:06:45] \u003c/em>Open AI, I mean you seem somewhat frustrated with them. You were one of the big contributors early on?\u003c/p>\n\u003cp>\u003cb>Elon Musk: \u003c/b>\u003cem>[00:06:49] \u003c/em>The reason, I am the reason Open AI exists.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:06:54] \u003c/em>So he wants more than his money back. He wants Altman and OpenAI’s co-founder and president, Greg Brockman, taken off the board. And he wants $130 billion, disgorged by the for-profit and handed over to the non-profit. The word charity, Ericka, doesn’t appear once in OpenAI’s founding blog post, but Musk keeps referring to OpenAI as a charity. But as OpenAI lawyers like to point out, Musk left OpenAI and then he launched his own AI venture, XAI, which is not a nonprofit and arguably does not operate for the benefit of humanity, for which it has been sued repeatedly.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:07:50] \u003c/em>So it sounds like Elon Musk is basically saying they stole his charity, and Sam Altman is saying, ‘You chose to walk away.’\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:08:02] \u003c/em>Yeah. That’s it in a nutshell. There was this funny moment when Musk was on the witness stand. He looked at the jury and he said, quote, it’s not OK to steal a charity. And then he predicted that if Open AI wins this case, the face of charity law in America could be altered forever. At some point, the judge broke in and said, let’s remind the jury, you’re not a lawyer. She’s talking to Musk. And then he replied. I did take Law 101, which got a laugh out of most people in the court.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:08:36] \u003c/em>Geez.\u003c/p>\n\u003cp>\u003cb>\u003c/b>\u003cem>[00:08:39] \u003c/em>Rachael, what do we make of Sam Altman’s role in this? It sounds like Elon Musk is saying that Sam Altman lied to him.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:08:46] \u003c/em>That is a very good question. I need to mention here that we have not seen Sam Altman take the stand in this trial yet. So Altman has not yet had the chance to make his case. Just a few weeks ago, we saw a comprehensive profile of Sam Altman in the New Yorker magazine talking to lots and lots of people that Sam Altmann is an inveterate liar, the kind of person who will tell you what you want to hear and then go back on it. We haven’t had the opportunity yet to really get into what his character was like during the early days of OpenAI, but pretty much everyone in that courtroom has read that article.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:09:47] \u003c/em>Coming up, what the OpenAI trial is really about. Stay with us.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:10:38] \u003c/em>I mean, Rachael, I gotta say, as I’m reading these stories about this case, it really just sounds like a fight between two of some of the richest billionaires in Silicon Valley over this company that they co-founded. But obviously, what’s at the center of it and what is at stake is this very powerful technology that even they seem to acknowledge has the potential to change the world. So what do you think this is really about?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:11:13] \u003c/em>Clearly about power, clearly about money, clearly about market dominance. And I do want to say that even though the judge is saying we are not going to talk about the AI apocalypse, it is something that is genuinely on the minds of all of these people in the industry in Silicon Valley and also the rest of us, right? I mean there are people here who take AI safety seriously. Who also think OpenAI has drifted dangerously from its mission. I mean, we’ve seen bad actors using the software who have upended the labor market, terrified all of us from a cybersecurity perspective, made it impossible to get redress as a customer and sometimes as a citizen, enabled a surveillance state here and abroad. I mean I could go on, Ericka, because It’s 100% clear to us and the people building this software that there’s a race to the bottom going on from a moral perspective.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:12:30] \u003c/em>I do want to ask you this question, Rachel, because Elon Musk is saying in this trial that he is the one standing up for the public on AI. Rachel, is there someone working in the public interest when it comes to AI and holding AI companies accountable?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:12:53] \u003c/em>Well, you know, don’t we wish? He’d like to present himself as thinking pro-human first, but you know, he also created XAI and has reportedly personally directed his engineers to make XAI a manifestly unsafe product. The judge noted the irony out loud. She said to Musk’s attorneys at one point, It is ironic that your client, despite these risks, is creating a company in the exact same space. And then she added, and I just thought this was so remarkable, coming from, again, a sitting federal judge, quote, I suspect there are people who don’t want to put the future in Mr. Musk’s hands, unquote.\u003c/p>\n\u003cp>\u003cb>Jill Horowitz: \u003c/b>\u003cem>[00:13:44] \u003c/em>And in that sense, I don’t understand why Musk is the one who gets to ask that question. Jill Horowitz, who specifically specializes in non-profit law. At Northwestern’s law school put it this way when parties have this much money and this much power they can trample over conventional protections of the public interest\u003c/p>\n\u003cp>\u003cb>Jill Horowitz: \u003c/b>\u003cem>[00:14:07] \u003c/em>We’ve got a CEO who is a very powerful player. And then we have this outside party who’s purported to be thinking about the best interest of the nonprofit, but he’s a competitor.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:14:22] \u003c/em>Congress, you don’t need me to tell you, hasn’t passed any meaningful federal AI regulation. The Trump administration is lobbying alongside the lobbyists for unfettered freedom for the AI industry. And so we end up here, Ericka, in a federal courthouse in Oakland watching two billionaires fight over their recent past. This trial gives us a window into the wheeling and dealing. But it doesn’t give us any power to change the trajectory of AI.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:14:54] \u003c/em>Rachael what happens if if either Elon Musk or Sam Altman wins this trial?\u003c/p>\n\u003cp>\u003cb>\u003c/b>\u003cem>[00:15:01] \u003c/em>So if Musk wins, Judge Gonzalez-Rogers could order OpenAI to revert to a non-profit structure, remove Altman and Brockman, direct some $130 billion in gains back to the non- profit foundation. That would be legally unprecedented and would certainly send shockwaves throughout Silicon Valley. If OpenAI wins, the restructuring stands, the IPO proceeds. And the message to the industry is essentially, you can do this too. You can take a non-profit, make it nominally in charge of a for-profit arm that you build into a trillion-dollar company, and the legal system won’t stop you.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:15:52] \u003c/em>Last question for you, Rachael. For the protesters outside, what do you think they want to see happen? And do you they care here about who wins?\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:16:03] \u003c/em>My sense from talking with Ashley Ortiz, who was one of the organizers of the first and biggest protest outside, is that for a lot of the people out here carrying signs and wearing t-shirts that say stop AI, neither Musk nor Altman represents their interests and by extension the public’s interests.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:16:28] \u003c/em>Decision everyone sucks here and y’all both need to take responsibility for your part in this crappy situation.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:16:34] \u003c/em>They want accountability for AI, period. I don’t know if they actually think they’re gonna get what they’re asking for, but they wanna make a noise while they can.\u003c/p>\n\u003cp>\u003cb>Ashley Ortiz: \u003c/b>\u003cem>[00:16:46] \u003c/em>We’re letting them both know that both sides, no matter which side wins, the people are going to lose because they are not doing this actually for the benefit of humanity. It’s not about ethics. This is all about power plays within an unfettered, unregulated AI scape.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:17:06] \u003c/em>These are the models that are changing our world, and they’re doing it now. And regardless of whether OpenAI survives this trial, we’re still gonna have the world that OpenAI helped to create.\u003c/p>\n\u003cp>\u003cb>Ericka Cruz Guevarra: \u003c/b>\u003cem>[00:17:28] \u003c/em>Well, Rachael, thank you so much for chatting with me outside the courtroom and for making the time in your busy morning, I appreciate it.\u003c/p>\n\u003cp>\u003cb>Rachael Myrow: \u003c/b>\u003cem>[00:17:36] \u003c/em>You bet.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "floatright"
},
"numeric": [
"floatright"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>\u003ci>\u003cspan style=\"font-weight: 400\">Some members of the KQED podcast team are represented by The Screen Actors Guild, American Federation of Television and Radio Artists, San Francisco-Northern California Local.\u003c/span>\u003c/i>\u003c/p>\n\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12082428/inside-sam-altman-and-elon-musks-battle-over-openai",
"authors": [
"8654",
"251",
"11649",
"11831"
],
"categories": [
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_32668",
"news_3897",
"news_36810",
"news_33812",
"news_35758",
"news_22598"
],
"featImg": "news_12082344",
"label": "source_news_12082428"
},
"news_12082064": {
"type": "posts",
"id": "news_12082064",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12082064",
"score": null,
"sort": [
1777676594000
]
},
"guestAuthors": [],
"slug": "openai-back-in-court-over-canada-school-shooters-use-of-chatgpt",
"title": "OpenAI Back in Court Over Canada School Shooter’s Use of ChatGPT",
"publishDate": 1777676594,
"format": "standard",
"headTitle": "OpenAI Back in Court Over Canada School Shooter’s Use of ChatGPT | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>The families of victims of a school shooting in a British Columbia town sued artificial intelligence company \u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI \u003c/a>in a San Francisco court this week, alleging that the company behind \u003ca href=\"https://www.kqed.org/news/tag/chatgpt\">ChatGPT\u003c/a> failed to alert police of the shooter’s alarming interactions with the chatbot.\u003c/p>\n\u003cp>One of the lawsuits was filed on behalf of Shannda Aviugana-Durand, an education assistant who was shot and killed in a library at \u003ca href=\"https://apnews.com/article/openai-chatgpt-tumbler-ridge-26d632a065e8420a263655c8ee29f961\">Tumbler Ridge Secondary School\u003c/a>. The suit alleges negligence, aiding and abetting a mass shooting, wrongful death and liability, among other claims. According to the lawsuit, Aviugana-Durand’s daughter was present at the time of the attack.\u003c/p>\n\u003cp>The educational assistant was one of six people who were killed by an 18-year-old in February. The teen — who later shot herself — also killed her mother and her 11-year-old half-brother at home beforehand. Twenty-five people were also injured in the attack, Canada’s deadliest mass shooting in years.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>Another lawsuit was filed Wednesday on behalf of 12-year-old Maya Gebala, who was critically injured in the February shooting. The plaintiffs’ attorney, Jay Edelson, said in an interview with the \u003cem>Associated Press\u003c/em> that decisions made by OpenAI and its CEO Sam Altman “have destroyed the town. The people are really resilient, but what happened is unimaginable.”\u003c/p>\n\u003cp>Altman sent a letter last week \u003ca href=\"https://apnews.com/article/openai-altman-tumbler-ridge-killings-apology-dec2adaad3946583519370eede6a99e2\">formally apologizing\u003c/a> to the community that his company did not notify law enforcement about the shooter’s online behavior in the weeks leading up to the attack.\u003c/p>\n\u003cp>The case highlights concerns about the harms posed by \u003ca href=\"https://apnews.com/article/ai-sycophancy-chatbots-science-study-8dc61e69278b661cab1e53d38b4173b6\">overly agreeable AI chatbots\u003c/a> and what obligations the tech industry has to control them or notify authorities about planned violence by chatbot users. This month, \u003ca href=\"https://apnews.com/article/missing-grad-students-florida-6279adeef3d0540865de39ab3d6f8093\">prosecutors investigating the deaths\u003c/a> of two University of South Florida doctoral students said that the suspect asked ChatGPT about body disposal in the lead-up to the students’ disappearance.\u003c/p>\n\u003cfigure id=\"attachment_12079761\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12079761 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman speaks during the BlackRock Infrastructure Summit on March 11, 2026, in Washington, D.C. \u003ccite>(Anna Moneymaker/Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“It’s not the first lawsuit of its kind,” said Robin Feldman, law professor at UC Law San Francisco and director of its AI Law and Innovation Institute. “This is part of an early wave of lawsuits in which citizens are asking to hold LLMs responsible for harms that happen down the line, whether they are crimes, mental health problems, suicide.”\u003c/p>\n\u003cp>“ChatGPT was first on the scene. And it is the most widely known of the LLMs,” Feldman said. “That puts it in the hot seat as the law tries to understand how to wrangle this unusual beast.”\u003c/p>\n\u003cp>In response to the lawsuit, OpenAI said in a written statement that the “events in Tumbler Ridge are a tragedy. We have a zero-tolerance policy for using our tools to assist in committing violence.”[aside postID=news_12081916 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/AP26118555622828-2000x1333.jpg']“As we shared with Canadian officials, we have already strengthened our safeguards, including improving how ChatGPT responds to signs of distress, connecting people with local support and mental health resources, strengthening how we assess and escalate potential threats of violence, and improving detection of repeat policy violators,” the company said.\u003c/p>\n\u003cp>Edelson, a Chicago-based lawyer known for taking on the tech industry, is already juggling a number of high-profile cases against OpenAI, including from the family of a California teenager who killed himself after \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">conversations with ChatGPT\u003c/a> and another from the heirs of an 83-year-old Connecticut woman \u003ca href=\"https://apnews.com/article/ai-chatgpt-wrongful-death-lawsuit-greenwich-97fd7da31c0fa08f3d3ea9efd6713151\">killed by her son\u003c/a> after ChatGPT allegedly amplified the man’s “paranoid delusions.”\u003c/p>\n\u003cp>“This is not a passive technology,” Edelson said, comparing the chatbot interactions with a more conventional online search for information. “What we’ve seen in the past is that (for) people who are mentally ill, the chatbot will validate what they’re saying and then amplify what they’re saying.”\u003c/p>\n\u003cp>Last week, Edelson visited the small town of Tumbler Ridge and met with dozens of people in the basement of a visitor center. He also visited Gebala at a children’s hospital in Vancouver, where she remains hospitalized and seemed alert but unable to speak.\u003c/p>\n\u003cp>“It was so heartbreaking,” he said.\u003c/p>\n\u003cfigure id=\"attachment_12082198\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082198 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Candles, flowers, photographs, plush toys and other items at a makeshift memorial for the victims four days after a deadly mass shooting took place at a school, in the town of Tumbler Ridge, British Columbia, Canada, on Feb. 13, 2026. \u003ccite>(Paige Taylor White/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>The lawsuits filed Wednesday also represent the families of the five slain children targeted in the school shooting: Zoey Benoit, Abel Mwansa Jr., Ticaria “Tiki” Lampert and Kylie Smith, all 12, and Ezekiel Schofield, 13.\u003c/p>\n\u003cp>After the shootings, OpenAI came forward to say that last June, the company flagged the shooter’s account as having been used to discuss violence against other people.\u003c/p>\n\u003cp>The company said it considered whether to refer the account to the Royal Canadian Mounted Police, but determined at the time that the account activity didn’t meet a threshold for referral to law enforcement. OpenAI banned the account in June for violating its usage policy.[aside postID=news_12080610 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1020x680.jpg']The lawsuits filed Wednesday allege “the victims didn’t learn this because OpenAI was forthcoming, but because \u003ca href=\"https://www.wsj.com/us-news/law/openai-employees-raised-alarms-about-canada-shooting-suspect-months-ago-b585df62\">its own employees leaked it to \u003cem>The Wall Street Journal\u003c/em>\u003c/a> after they could no longer stomach the company’s silence.”\u003c/p>\n\u003cp>In \u003ca href=\"https://tumblerridgelines.com/2026/04/24/openai-apologizes-to-tumbler-ridge/\">his letter\u003c/a>, Altman said he was “deeply sorry that we did not alert law enforcement to the account that was banned in June.”\u003c/p>\n\u003cp>“While I know words can never be enough, I believe an apology is necessary to recognize the harm and irreversible loss your community has suffered,” Altman wrote.\u003c/p>\n\u003cp>British Columbia Premier David Eby, \u003ca href=\"https://x.com/dave_eby/status/2047751590803886291?s=46&t=7BBzFwo6eYLzJIVfAlumEQ\">in a social media post\u003c/a>, called the apology “necessary, and yet grossly insufficient for the devastation done to the families of Tumbler Ridge.”\u003c/p>\n\u003cp>The Gebala lawsuit accuses OpenAI of negligence involving a failure to warn law enforcement and “aiding and abetting a mass shooting.”\u003c/p>\n\u003cp>Along with damages, the Gebala lawsuit seeks a court order that would require OpenAI to ban users from ChatGPT if their accounts were deactivated for violent misuse, and to require the company to alert law enforcement when its systems identify someone who poses a “real-world risk of violence.”\u003c/p>\n\u003cp>An earlier case was filed in a court in British Columbia, but a team of lawyers in both countries is seeking to bring the affiliated cases to San Francisco, where OpenAI is headquartered.\u003c/p>\n\u003ch2>‘Untried territory’\u003c/h2>\n\u003cp>Feldman called reports that the company flagged the risk but failed to act effectively “deeply troubling.”\u003c/p>\n\u003cp>“As with so much about AI, the lawsuit will take us into untried territory,” she said. “The old doctrines are being applied to new circumstances.”\u003c/p>\n\u003cp>She said if the families were to win, the company would have to pay damages and assume responsibility for altering its platform to identify and respond to risks.\u003c/p>\n\u003cp>The major issues that the lawsuit will tackle are whether OpenAI and ChatGPT are protected by the First Amendment and whether or not OpenAI had “a duty to act,” she said.\u003c/p>\n\u003cfigure id=\"attachment_12082201\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082201 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Community members attend a vigil to honor the victims of one of Canada’s deadliest mass shootings in Tumbler Ridge, British Columbia, Canada, on Feb. 13, 2026. \u003ccite>(Paige Taylor White/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>She said that there are \u003ca href=\"https://www.congress.gov/crs-product/R46751\">parts\u003c/a> of U.S. law that shield tech companies from liability for content that their users host. Essentially, this means platforms are more like “bulletin boards” and “are not responsible for the content.”\u003c/p>\n\u003cp>But this case would raise the question, she said, “Are LLMs like a bulletin board or publisher? Or they like a facilitator who helped with the crime?”\u003c/p>\n\u003cp>Some companies struggle with the burden of responsibility when reviewing potential threats to public safety, Feldman said, “If they try to help out, they can be viewed as accepting the mantle of responsibility.”\u003c/p>\n\u003cp>According to Feldman, families are also likely to argue that the LLM “is a defective product without appropriate safeguards.\u003c/p>\n\u003cp>“In that case, the question is the following: ‘Is the LLM a defective product, or merely a product that was used improperly? And is it analogous to a product at all?”\u003c/p>\n\u003cp>“All of these are tough questions as we enter the age of AI, and the courts are just beginning to explore them,” Feldman said.\u003c/p>\n\u003cp>\u003cem>The Associated Press’ Jim Morris contributed to this story.\u003c/em>\u003c/p>\n\u003cp>[ad floatright]\u003c/p>\n",
"blocks": [],
"excerpt": "The lawsuit alleges negligence and wrongful death on account of the shooter’s interactions with the chatbot in the weeks and months leading up to the fatal attack.",
"status": "publish",
"parent": 0,
"modified": 1787327550,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 35,
"wordCount": 1495
},
"headData": {
"title": "OpenAI Back in Court Over Canada School Shooter’s Use of ChatGPT | KQED",
"description": "The lawsuit alleges negligence and wrongful death on account of the shooter’s interactions with the chatbot in the weeks and months leading up to the fatal attack.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "OpenAI Back in Court Over Canada School Shooter’s Use of ChatGPT",
"datePublished": "2026-05-01T16:03:14-07:00",
"dateModified": "2026-08-21T08:52:30-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 34167,
"slug": "criminal-justice",
"name": "Criminal Justice"
},
"sticky": false,
"nprByline": "Matt O’Brien, Associated Press, and Nisa Khan, KQED",
"nprStoryId": "kqed-12082064",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"showOnAuthorArchivePages": "Yes",
"articleAge": "0",
"path": "/news/12082064/openai-back-in-court-over-canada-school-shooters-use-of-chatgpt",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>The families of victims of a school shooting in a British Columbia town sued artificial intelligence company \u003ca href=\"https://www.kqed.org/news/tag/open-ai\">OpenAI \u003c/a>in a San Francisco court this week, alleging that the company behind \u003ca href=\"https://www.kqed.org/news/tag/chatgpt\">ChatGPT\u003c/a> failed to alert police of the shooter’s alarming interactions with the chatbot.\u003c/p>\n\u003cp>One of the lawsuits was filed on behalf of Shannda Aviugana-Durand, an education assistant who was shot and killed in a library at \u003ca href=\"https://apnews.com/article/openai-chatgpt-tumbler-ridge-26d632a065e8420a263655c8ee29f961\">Tumbler Ridge Secondary School\u003c/a>. The suit alleges negligence, aiding and abetting a mass shooting, wrongful death and liability, among other claims. According to the lawsuit, Aviugana-Durand’s daughter was present at the time of the attack.\u003c/p>\n\u003cp>The educational assistant was one of six people who were killed by an 18-year-old in February. The teen — who later shot herself — also killed her mother and her 11-year-old half-brother at home beforehand. Twenty-five people were also injured in the attack, Canada’s deadliest mass shooting in years.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>Another lawsuit was filed Wednesday on behalf of 12-year-old Maya Gebala, who was critically injured in the February shooting. The plaintiffs’ attorney, Jay Edelson, said in an interview with the \u003cem>Associated Press\u003c/em> that decisions made by OpenAI and its CEO Sam Altman “have destroyed the town. The people are really resilient, but what happened is unimaginable.”\u003c/p>\n\u003cp>Altman sent a letter last week \u003ca href=\"https://apnews.com/article/openai-altman-tumbler-ridge-killings-apology-dec2adaad3946583519370eede6a99e2\">formally apologizing\u003c/a> to the community that his company did not notify law enforcement about the shooter’s online behavior in the weeks leading up to the attack.\u003c/p>\n\u003cp>The case highlights concerns about the harms posed by \u003ca href=\"https://apnews.com/article/ai-sycophancy-chatbots-science-study-8dc61e69278b661cab1e53d38b4173b6\">overly agreeable AI chatbots\u003c/a> and what obligations the tech industry has to control them or notify authorities about planned violence by chatbot users. This month, \u003ca href=\"https://apnews.com/article/missing-grad-students-florida-6279adeef3d0540865de39ab3d6f8093\">prosecutors investigating the deaths\u003c/a> of two University of South Florida doctoral students said that the suspect asked ChatGPT about body disposal in the lead-up to the students’ disappearance.\u003c/p>\n\u003cfigure id=\"attachment_12079761\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12079761 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/SamAltmanGetty2-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI CEO Sam Altman speaks during the BlackRock Infrastructure Summit on March 11, 2026, in Washington, D.C. \u003ccite>(Anna Moneymaker/Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>“It’s not the first lawsuit of its kind,” said Robin Feldman, law professor at UC Law San Francisco and director of its AI Law and Innovation Institute. “This is part of an early wave of lawsuits in which citizens are asking to hold LLMs responsible for harms that happen down the line, whether they are crimes, mental health problems, suicide.”\u003c/p>\n\u003cp>“ChatGPT was first on the scene. And it is the most widely known of the LLMs,” Feldman said. “That puts it in the hot seat as the law tries to understand how to wrangle this unusual beast.”\u003c/p>\n\u003cp>In response to the lawsuit, OpenAI said in a written statement that the “events in Tumbler Ridge are a tragedy. We have a zero-tolerance policy for using our tools to assist in committing violence.”\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12081916",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/AP26118555622828-2000x1333.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>“As we shared with Canadian officials, we have already strengthened our safeguards, including improving how ChatGPT responds to signs of distress, connecting people with local support and mental health resources, strengthening how we assess and escalate potential threats of violence, and improving detection of repeat policy violators,” the company said.\u003c/p>\n\u003cp>Edelson, a Chicago-based lawyer known for taking on the tech industry, is already juggling a number of high-profile cases against OpenAI, including from the family of a California teenager who killed himself after \u003ca href=\"https://apnews.com/article/ai-chatbot-teens-congress-chatgpt-character-ce3959b6a3ea1a4997bf1ccabb4f0de2\">conversations with ChatGPT\u003c/a> and another from the heirs of an 83-year-old Connecticut woman \u003ca href=\"https://apnews.com/article/ai-chatgpt-wrongful-death-lawsuit-greenwich-97fd7da31c0fa08f3d3ea9efd6713151\">killed by her son\u003c/a> after ChatGPT allegedly amplified the man’s “paranoid delusions.”\u003c/p>\n\u003cp>“This is not a passive technology,” Edelson said, comparing the chatbot interactions with a more conventional online search for information. “What we’ve seen in the past is that (for) people who are mentally ill, the chatbot will validate what they’re saying and then amplify what they’re saying.”\u003c/p>\n\u003cp>Last week, Edelson visited the small town of Tumbler Ridge and met with dozens of people in the basement of a visitor center. He also visited Gebala at a children’s hospital in Vancouver, where she remains hospitalized and seemed alert but unable to speak.\u003c/p>\n\u003cp>“It was so heartbreaking,” he said.\u003c/p>\n\u003cfigure id=\"attachment_12082198\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082198 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty2-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Candles, flowers, photographs, plush toys and other items at a makeshift memorial for the victims four days after a deadly mass shooting took place at a school, in the town of Tumbler Ridge, British Columbia, Canada, on Feb. 13, 2026. \u003ccite>(Paige Taylor White/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>The lawsuits filed Wednesday also represent the families of the five slain children targeted in the school shooting: Zoey Benoit, Abel Mwansa Jr., Ticaria “Tiki” Lampert and Kylie Smith, all 12, and Ezekiel Schofield, 13.\u003c/p>\n\u003cp>After the shootings, OpenAI came forward to say that last June, the company flagged the shooter’s account as having been used to discuss violence against other people.\u003c/p>\n\u003cp>The company said it considered whether to refer the account to the Royal Canadian Mounted Police, but determined at the time that the account activity didn’t meet a threshold for referral to law enforcement. OpenAI banned the account in June for violating its usage policy.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12080610",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2024/06/GettyImages-2155035557-1020x680.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>The lawsuits filed Wednesday allege “the victims didn’t learn this because OpenAI was forthcoming, but because \u003ca href=\"https://www.wsj.com/us-news/law/openai-employees-raised-alarms-about-canada-shooting-suspect-months-ago-b585df62\">its own employees leaked it to \u003cem>The Wall Street Journal\u003c/em>\u003c/a> after they could no longer stomach the company’s silence.”\u003c/p>\n\u003cp>In \u003ca href=\"https://tumblerridgelines.com/2026/04/24/openai-apologizes-to-tumbler-ridge/\">his letter\u003c/a>, Altman said he was “deeply sorry that we did not alert law enforcement to the account that was banned in June.”\u003c/p>\n\u003cp>“While I know words can never be enough, I believe an apology is necessary to recognize the harm and irreversible loss your community has suffered,” Altman wrote.\u003c/p>\n\u003cp>British Columbia Premier David Eby, \u003ca href=\"https://x.com/dave_eby/status/2047751590803886291?s=46&t=7BBzFwo6eYLzJIVfAlumEQ\">in a social media post\u003c/a>, called the apology “necessary, and yet grossly insufficient for the devastation done to the families of Tumbler Ridge.”\u003c/p>\n\u003cp>The Gebala lawsuit accuses OpenAI of negligence involving a failure to warn law enforcement and “aiding and abetting a mass shooting.”\u003c/p>\n\u003cp>Along with damages, the Gebala lawsuit seeks a court order that would require OpenAI to ban users from ChatGPT if their accounts were deactivated for violent misuse, and to require the company to alert law enforcement when its systems identify someone who poses a “real-world risk of violence.”\u003c/p>\n\u003cp>An earlier case was filed in a court in British Columbia, but a team of lawyers in both countries is seeking to bring the affiliated cases to San Francisco, where OpenAI is headquartered.\u003c/p>\n\u003ch2>‘Untried territory’\u003c/h2>\n\u003cp>Feldman called reports that the company flagged the risk but failed to act effectively “deeply troubling.”\u003c/p>\n\u003cp>“As with so much about AI, the lawsuit will take us into untried territory,” she said. “The old doctrines are being applied to new circumstances.”\u003c/p>\n\u003cp>She said if the families were to win, the company would have to pay damages and assume responsibility for altering its platform to identify and respond to risks.\u003c/p>\n\u003cp>The major issues that the lawsuit will tackle are whether OpenAI and ChatGPT are protected by the First Amendment and whether or not OpenAI had “a duty to act,” she said.\u003c/p>\n\u003cfigure id=\"attachment_12082201\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"wp-image-12082201 size-full\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3.jpg\" alt=\"\" width=\"2000\" height=\"1333\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3-160x107.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/05/TumblerRidgeGetty3-1536x1024.jpg 1536w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Community members attend a vigil to honor the victims of one of Canada’s deadliest mass shootings in Tumbler Ridge, British Columbia, Canada, on Feb. 13, 2026. \u003ccite>(Paige Taylor White/AFP via Getty Images)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>She said that there are \u003ca href=\"https://www.congress.gov/crs-product/R46751\">parts\u003c/a> of U.S. law that shield tech companies from liability for content that their users host. Essentially, this means platforms are more like “bulletin boards” and “are not responsible for the content.”\u003c/p>\n\u003cp>But this case would raise the question, she said, “Are LLMs like a bulletin board or publisher? Or they like a facilitator who helped with the crime?”\u003c/p>\n\u003cp>Some companies struggle with the burden of responsibility when reviewing potential threats to public safety, Feldman said, “If they try to help out, they can be viewed as accepting the mantle of responsibility.”\u003c/p>\n\u003cp>According to Feldman, families are also likely to argue that the LLM “is a defective product without appropriate safeguards.\u003c/p>\n\u003cp>“In that case, the question is the following: ‘Is the LLM a defective product, or merely a product that was used improperly? And is it analogous to a product at all?”\u003c/p>\n\u003cp>“All of these are tough questions as we enter the age of AI, and the courts are just beginning to explore them,” Feldman said.\u003c/p>\n\u003cp>\u003cem>The Associated Press’ Jim Morris contributed to this story.\u003c/em>\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "floatright"
},
"numeric": [
"floatright"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12082064/openai-back-in-court-over-canada-school-shooters-use-of-chatgpt",
"authors": [
"byline_news_12082064"
],
"categories": [
"news_34167",
"news_28250",
"news_8"
],
"tags": [
"news_34755",
"news_1386",
"news_32668",
"news_17725",
"news_22434",
"news_35784",
"news_33542",
"news_33543",
"news_38"
],
"featImg": "news_12082068",
"label": "news"
},
"news_12081798": {
"type": "posts",
"id": "news_12081798",
"meta": {
"index": "posts_1716263798",
"site": "news",
"id": "12081798",
"score": null,
"sort": [
1777507270000
]
},
"guestAuthors": [],
"slug": "elon-musk-says-sam-altman-tricked-him-into-funding-openai",
"title": "Elon Musk Says Sam Altman Tricked Him Into Funding OpenAI",
"publishDate": 1777507270,
"format": "standard",
"headTitle": "Elon Musk Says Sam Altman Tricked Him Into Funding OpenAI | KQED",
"labelTerm": {
"site": "news"
},
"content": "\u003cp>During the second day of the \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">landmark trial between Sam Altman and Elon Musk\u003c/a>, the Tesla founder told the Oakland courthouse that he was a “fool” to fund OpenAI through its early years.\u003c/p>\n\u003cp>Testifying in the lawsuit he brought against Altman, which claims the company’s creators betrayed their mission for profits, Musk suggested Wednesday that Altman and cofounder Greg Brockman wanted to “have your cake and eat it too.”\u003c/p>\n\u003cp>“If you go nonprofit, you’ve got a sort of moral high ground,” he testified.\u003c/p>\n\u003cp>Musk’s testimony tells one version of founding OpenAI: that he, fearing the dangers of artificial intelligence, pursued its development with the goal of benefiting the common good, alongside, he thought, like-minded collaborators. But behind the scenes, those cofounders engaged in a “long con” to profit at his expense.\u003c/p>\n\u003cp>[ad fullwidth]\u003c/p>\n\u003cp>“What they really wanted was a for-profit, where they could make as much money as possible,” Musk said later.\u003c/p>\n\u003cp>Whether the jury believes him will be integral to the decision they’re tasked with making, as they determine whether OpenAI breached charitable trust and engaged in unjust enrichment as it evolved from a nonprofit organization to its current $730 billion iteration.\u003c/p>\n\u003cp>Under cross-examination, Altman’s attorney, William Savitt, questioned Musk’s story and credibility as an altruistic benefactor. He pointed to an email Musk sent to Altman in 2015, which said it would be “probably better” if OpenAI operated as a for-profit company with a parallel nonprofit.\u003c/p>\n\u003cfigure id=\"attachment_12081637\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081637\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI’s lead counsel, William Savitt, presents opening statements in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>In another email sent to colleagues at his neurotechnology company, Neuralink, Musk said that Google’s AI development was moving very fast, and that he was concerned OpenAI was not on the path to catch up.\u003c/p>\n\u003cp>“Setting it up as a nonprofit might, in hindsight, have been the wrong move,” Musk wrote. “Sense of urgency is not as high.”\u003c/p>\n\u003cp>Savitt asked if, in 2017, Musk suggested at a party that OpenAI should create a for-profit. He said it was just after the company’s AI model had beaten \u003cem>Defense of the Ancients, \u003c/em>a battle video game, which was a pivotal moment in the development process.\u003c/p>\n\u003cp>Musk said he didn’t remember giving instructions to create a for-profit at the time.\u003c/p>\n\u003cp>“This was nine years ago,” he said.[aside postID=news_12081603 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-02-KQED.jpg']Savitt said Tuesday that in 2017, OpenAI executives, including Musk, were in the midst of conversations about whether and how to transition the company to a for-profit structure.\u003c/p>\n\u003cp>According to OpenAI’s court filings, as early as summer 2017, Musk had insisted on holding a majority equity stake in any for-profit entity, serving as CEO and controlling its board of directors.\u003c/p>\n\u003cp>Pressed by Savitt about what Musk meant by “expressing what you said about control,” the Tesla founder and billionaire said: “I try to be as literal as possible.”\u003c/p>\n\u003cp>In the fall of 2017, Brockman and Ilya Sutskever, another top OpenAI executive, emailed Musk with concerns about the for-profit structure he proposed. Shortly thereafter, discussions over the structure collapsed, and Musk stopped making significant quarterly funding contributions, OpenAI alleges.\u003c/p>\n\u003cp>He left the company less than six months later.\u003c/p>\n\u003cp>Savitt framed the breakdown and Musk’s exit as a result of his not getting control of the for-profit, and the other executives’ focus on maintaining its philanthropic mission. He suggested that Musk tried to pressure them to accept his terms by pausing the majority of his financial backing.\u003c/p>\n\u003cp>“You knew that would create financial pressure for the organization,” Savitt said.\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081686\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Musk denied that was his intention. Instead, he alleged that Altman convinced Brockman and the others to go against his proposal, and that their concern over his desire for control was disingenuous.\u003c/p>\n\u003cp>“I’m not going to fund something if I don’t have confidence in the people,” he said.\u003c/p>\n\u003cp>When asked whether he proposed that OpenAI be folded into Tesla, Musk said: “There were a lot of ideas that were brainstormed at the time.”\u003c/p>\n\u003cp>In an email, he wrote that doing so would be the “only path that could even hope to hold a candle to Google.”\u003c/p>\n\u003cp>Musk said he left OpenAI in February 2018 because he was focused on Tesla’s survival, and believed that OpenAI intended to continue operating as a nonprofit.\u003c/p>\n\u003cp>Savitt also laid out a series of exchanges between Musk and Altman, in which the OpenAI CEO kept him apprised of the company’s corporate structure. He said in March 2018, Musk responded to an email that noted the creation of a for-profit entity of OpenAI with “OK by me,” and was sent a term sheet for OpenAI LP that summer.[aside postID=news_12081290 hero='https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260422-ALTMANMUSK-MD-01-KQED.jpg']Savitt also said Altman emailed Musk a draft of the company’s public announcement of its for-profit arm in March 2019, and texted him asking if he had time to talk about Microsoft’s plan to invest in OpenAI. Musk never responded to that text, according to Savitt.\u003c/p>\n\u003cp>Musk said he was busy with his other companies in 2018, and while he was aware that it had added a for-profit entity, he hadn’t lost complete faith in the company. While he’d suspended quarterly $5 million funding contributions prior to his departure, he continued to make some contributions until 2020.\u003c/p>\n\u003cp>He said that he’d gone from enthusiastically supportive to uncertain about OpenAI’s mission, but that he’d fully suspended his contributions when he felt that the company was “deliberately not a nonprofit.”\u003c/p>\n\u003cp>When asked why he waited until 2024 to bring the suit, Musk said that’s when he determined OpenAI breached charitable trust.\u003c/p>\n\u003cp>“Thinking that someone might steal your car is not the same as [if] someone has stolen your car,” Musk said. He said after enlisting his attorney, Alex Spiro, to investigate, he heard from him in 2023 that “the car had been stolen.”\u003c/p>\n\u003cp>“I would have sued sooner if I thought the charity had been stolen sooner,” Musk continued.\u003c/p>\n\u003cp>The trial and Musk’s testimony are expected to continue on Thursday.\u003c/p>\n\u003cp>[ad floatright]\u003c/p>\n",
"blocks": [],
"excerpt": "On the second day of a trial pitting the Tesla founder against OpenAI, Elon Musk said he was a “fool” to support the company behind ChatGPT during its early years.",
"status": "publish",
"parent": 0,
"modified": 1777509912,
"stats": {
"hasAudio": false,
"hasVideo": false,
"hasChartOrMap": false,
"iframeSrcs": [],
"hasGoogleForm": false,
"hasGallery": false,
"hasHearkenModule": false,
"hasPolis": false,
"paragraphCount": 32,
"wordCount": 1208
},
"headData": {
"title": "Elon Musk Says Sam Altman Tricked Him Into Funding OpenAI | KQED",
"description": "On the second day of a trial pitting the Tesla founder against OpenAI, Elon Musk said he was a “fool” to support the company behind ChatGPT during its early years.",
"ogTitle": "",
"ogDescription": "",
"ogImgId": "",
"twTitle": "",
"twDescription": "",
"twImgId": "",
"schema": {
"@context": "https://schema.org",
"@type": "NewsArticle",
"headline": "Elon Musk Says Sam Altman Tricked Him Into Funding OpenAI",
"datePublished": "2026-04-29T17:01:10-07:00",
"dateModified": "2026-04-29T17:45:12-07:00",
"image": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"isAccessibleForFree": "True",
"publisher": {
"@type": "NewsMediaOrganization",
"@id": "https://www.kqed.org/#organization",
"name": "KQED",
"logo": "https://cdn.kqed.org/wp-content/uploads/2020/02/KQED-OG-Image@1x.png",
"url": "https://www.kqed.org",
"sameAs": [
"https://www.facebook.com/KQED",
"https://twitter.com/KQED",
"https://www.instagram.com/kqed/",
"https://www.tiktok.com/@kqedofficial",
"https://www.linkedin.com/company/kqed",
"https://www.youtube.com/channel/UCeC0IOo7i1P_61zVUWbJ4nw"
]
}
}
},
"primaryCategory": {
"termId": 248,
"slug": "technology",
"name": "Technology"
},
"sticky": false,
"nprStoryId": "kqed-12081798",
"templateType": "standard",
"featuredImageType": "standard",
"excludeFromSiteSearch": "Include",
"articleAge": "0",
"path": "/news/12081798/elon-musk-says-sam-altman-tricked-him-into-funding-openai",
"audioTrackLength": null,
"parsedContent": [
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003cp>During the second day of the \u003ca href=\"https://www.kqed.org/news/12081603/elon-musk-takes-aim-at-openai-as-trial-begins-its-not-ok-to-steal-a-charity\">landmark trial between Sam Altman and Elon Musk\u003c/a>, the Tesla founder told the Oakland courthouse that he was a “fool” to fund OpenAI through its early years.\u003c/p>\n\u003cp>Testifying in the lawsuit he brought against Altman, which claims the company’s creators betrayed their mission for profits, Musk suggested Wednesday that Altman and cofounder Greg Brockman wanted to “have your cake and eat it too.”\u003c/p>\n\u003cp>“If you go nonprofit, you’ve got a sort of moral high ground,” he testified.\u003c/p>\n\u003cp>Musk’s testimony tells one version of founding OpenAI: that he, fearing the dangers of artificial intelligence, pursued its development with the goal of benefiting the common good, alongside, he thought, like-minded collaborators. But behind the scenes, those cofounders engaged in a “long con” to profit at his expense.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "fullwidth"
},
"numeric": [
"fullwidth"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003cp>“What they really wanted was a for-profit, where they could make as much money as possible,” Musk said later.\u003c/p>\n\u003cp>Whether the jury believes him will be integral to the decision they’re tasked with making, as they determine whether OpenAI breached charitable trust and engaged in unjust enrichment as it evolved from a nonprofit organization to its current $730 billion iteration.\u003c/p>\n\u003cp>Under cross-examination, Altman’s attorney, William Savitt, questioned Musk’s story and credibility as an altruistic benefactor. He pointed to an email Musk sent to Altman in 2015, which said it would be “probably better” if OpenAI operated as a for-profit company with a parallel nonprofit.\u003c/p>\n\u003cfigure id=\"attachment_12081637\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081637\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-01-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">OpenAI’s lead counsel, William Savitt, presents opening statements in the trial in which Elon Musk claims that Sam Altman and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>In another email sent to colleagues at his neurotechnology company, Neuralink, Musk said that Google’s AI development was moving very fast, and that he was concerned OpenAI was not on the path to catch up.\u003c/p>\n\u003cp>“Setting it up as a nonprofit might, in hindsight, have been the wrong move,” Musk wrote. “Sense of urgency is not as high.”\u003c/p>\n\u003cp>Savitt asked if, in 2017, Musk suggested at a party that OpenAI should create a for-profit. He said it was just after the company’s AI model had beaten \u003cem>Defense of the Ancients, \u003c/em>a battle video game, which was a pivotal moment in the development process.\u003c/p>\n\u003cp>Musk said he didn’t remember giving instructions to create a for-profit at the time.\u003c/p>\n\u003cp>“This was nine years ago,” he said.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12081603",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-02-KQED.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Savitt said Tuesday that in 2017, OpenAI executives, including Musk, were in the midst of conversations about whether and how to transition the company to a for-profit structure.\u003c/p>\n\u003cp>According to OpenAI’s court filings, as early as summer 2017, Musk had insisted on holding a majority equity stake in any for-profit entity, serving as CEO and controlling its board of directors.\u003c/p>\n\u003cp>Pressed by Savitt about what Musk meant by “expressing what you said about control,” the Tesla founder and billionaire said: “I try to be as literal as possible.”\u003c/p>\n\u003cp>In the fall of 2017, Brockman and Ilya Sutskever, another top OpenAI executive, emailed Musk with concerns about the for-profit structure he proposed. Shortly thereafter, discussions over the structure collapsed, and Musk stopped making significant quarterly funding contributions, OpenAI alleges.\u003c/p>\n\u003cp>He left the company less than six months later.\u003c/p>\n\u003cp>Savitt framed the breakdown and Musk’s exit as a result of his not getting control of the for-profit, and the other executives’ focus on maintaining its philanthropic mission. He suggested that Musk tried to pressure them to accept his terms by pausing the majority of his financial backing.\u003c/p>\n\u003cp>“You knew that would create financial pressure for the organization,” Savitt said.\u003c/p>\n\u003cfigure id=\"attachment_12081686\" class=\"wp-caption aligncenter\" style=\"max-width: 2000px\">\u003cimg loading=\"lazy\" decoding=\"async\" class=\"size-full wp-image-12081686\" src=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg\" alt=\"\" width=\"2000\" height=\"1125\" srcset=\"https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED.jpg 2000w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-160x90.jpg 160w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1536x864.jpg 1536w, https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260428-MUSK-ALTMAN-VB-04-KQED-1200x675.jpg 1200w\" sizes=\"auto, (max-width: 2000px) 100vw, 2000px\">\u003cfigcaption class=\"wp-caption-text\">Steve Molo, Elon Musk’s attorney, presents opening statements in the trial in which Elon Musk (center-right) claims that Sam Altman (right) and OpenAI abandoned their founding promise to develop AI for the benefit of humanity, rather than solely for profit, in Oakland on April 28, 2026. \u003ccite>(Vicki Behringer for KQED)\u003c/cite>\u003c/figcaption>\u003c/figure>\n\u003cp>Musk denied that was his intention. Instead, he alleged that Altman convinced Brockman and the others to go against his proposal, and that their concern over his desire for control was disingenuous.\u003c/p>\n\u003cp>“I’m not going to fund something if I don’t have confidence in the people,” he said.\u003c/p>\n\u003cp>When asked whether he proposed that OpenAI be folded into Tesla, Musk said: “There were a lot of ideas that were brainstormed at the time.”\u003c/p>\n\u003cp>In an email, he wrote that doing so would be the “only path that could even hope to hold a candle to Google.”\u003c/p>\n\u003cp>Musk said he left OpenAI in February 2018 because he was focused on Tesla’s survival, and believed that OpenAI intended to continue operating as a nonprofit.\u003c/p>\n\u003cp>Savitt also laid out a series of exchanges between Musk and Altman, in which the OpenAI CEO kept him apprised of the company’s corporate structure. He said in March 2018, Musk responded to an email that noted the creation of a for-profit entity of OpenAI with “OK by me,” and was sent a term sheet for OpenAI LP that summer.\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "aside",
"attributes": {
"named": {
"postid": "news_12081290",
"hero": "https://cdn.kqed.org/wp-content/uploads/sites/10/2026/04/260422-ALTMANMUSK-MD-01-KQED.jpg",
"label": ""
},
"numeric": []
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>Savitt also said Altman emailed Musk a draft of the company’s public announcement of its for-profit arm in March 2019, and texted him asking if he had time to talk about Microsoft’s plan to invest in OpenAI. Musk never responded to that text, according to Savitt.\u003c/p>\n\u003cp>Musk said he was busy with his other companies in 2018, and while he was aware that it had added a for-profit entity, he hadn’t lost complete faith in the company. While he’d suspended quarterly $5 million funding contributions prior to his departure, he continued to make some contributions until 2020.\u003c/p>\n\u003cp>He said that he’d gone from enthusiastically supportive to uncertain about OpenAI’s mission, but that he’d fully suspended his contributions when he felt that the company was “deliberately not a nonprofit.”\u003c/p>\n\u003cp>When asked why he waited until 2024 to bring the suit, Musk said that’s when he determined OpenAI breached charitable trust.\u003c/p>\n\u003cp>“Thinking that someone might steal your car is not the same as [if] someone has stolen your car,” Musk said. He said after enlisting his attorney, Alex Spiro, to investigate, he heard from him in 2023 that “the car had been stolen.”\u003c/p>\n\u003cp>“I would have sued sooner if I thought the charity had been stolen sooner,” Musk continued.\u003c/p>\n\u003cp>The trial and Musk’s testimony are expected to continue on Thursday.\u003c/p>\n\u003cp>\u003c/p>\u003c/div>",
"attributes": {
"named": {},
"numeric": []
}
},
{
"type": "component",
"content": "",
"name": "ad",
"attributes": {
"named": {
"label": "floatright"
},
"numeric": [
"floatright"
]
}
},
{
"type": "contentString",
"content": "\u003cdiv class=\"post-body\">\u003cp>\u003c/p>\n\u003c/div>\u003c/p>",
"attributes": {
"named": {},
"numeric": []
}
}
],
"link": "/news/12081798/elon-musk-says-sam-altman-tricked-him-into-funding-openai",
"authors": [
"11913",
"251"
],
"categories": [
"news_31795",
"news_6188",
"news_8",
"news_248"
],
"tags": [
"news_34755",
"news_32668",
"news_3897",
"news_27626",
"news_19954",
"news_21891",
"news_34054",
"news_33542",
"news_33543",
"news_34586",
"news_1631",
"news_57"
],
"featImg": "news_12081681",
"label": "news"
}
},
"podcastsReducer": {
"isFetching": false,
"fetchFailed": false,
"hasFetched": false,
"podcasts": {}
},
"radioProgramsReducer": {
"isFetching": false,
"fetchFailed": false,
"hasFetched": false,
"radioPrograms": {}
},
"radioSearchSegmentsReducer": {},
"programsReducer": {
"all-things-considered": {
"id": "all-things-considered",
"title": "All Things Considered",
"info": "Every weekday, \u003cem>All Things Considered\u003c/em> hosts Robert Siegel, Audie Cornish, Ari Shapiro, and Kelly McEvers present the program's trademark mix of news, interviews, commentaries, reviews, and offbeat features. Michel Martin hosts on the weekends.",
"airtime": "MON-FRI 1pm-2pm, 4:30pm-6:30pm\u003cbr />SAT-SUN 5pm-6pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/All-Things-Considered-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/all-things-considered/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/all-things-considered"
},
"american-suburb-podcast": {
"id": "american-suburb-podcast",
"title": "American Suburb: The Podcast",
"tagline": "The flip side of gentrification, told through one town",
"info": "Gentrification is changing cities across America, forcing people from neighborhoods they have long called home. Call them the displaced. Now those priced out of the Bay Area are looking for a better life in an unlikely place. American Suburb follows this migration to one California town along the Delta, 45 miles from San Francisco. But is this once sleepy suburb ready for them?",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/American-Suburb-Podcast-Tile-703x703-1.jpg",
"officialWebsiteLink": "/news/series/american-suburb-podcast",
"meta": {
"site": "news",
"source": "kqed",
"order": 19
},
"link": "/news/series/american-suburb-podcast/",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/RBrW",
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?mt=2&id=1287748328",
"tuneIn": "https://tunein.com/radio/American-Suburb-p1086805/",
"rss": "https://ww2.kqed.org/news/series/american-suburb-podcast/feed/podcast",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vS1FJTkMzMDExODgxNjA5"
}
},
"baycurious": {
"id": "baycurious",
"title": "Bay Curious",
"tagline": "Exploring the Bay Area, one question at a time",
"info": "KQED’s new podcast, Bay Curious, gets to the bottom of the mysteries — both profound and peculiar — that give the Bay Area its unique identity. And we’ll do it with your help! You ask the questions. You decide what Bay Curious investigates. And you join us on the journey to find the answers.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Bay-Curious-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Bay Curious",
"officialWebsiteLink": "/news/series/baycurious",
"meta": {
"site": "news",
"source": "kqed",
"order": 3
},
"link": "/podcasts/baycurious",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/bay-curious/id1172473406",
"npr": "https://www.npr.org/podcasts/500557090/bay-curious",
"rss": "https://ww2.kqed.org/news/category/bay-curious-podcast/feed/podcast",
"amazon": "https://music.amazon.com/podcasts/9a90d476-aa04-455d-9a4c-0871ed6216d4/bay-curious",
"stitcher": "https://www.stitcher.com/podcast/kqed/bay-curious",
"spotify": "https://open.spotify.com/show/6O76IdmhixfijmhTZLIJ8k"
}
},
"bbc-world-service": {
"id": "bbc-world-service",
"title": "BBC World Service",
"info": "The day's top stories from BBC News compiled twice daily in the week, once at weekends.",
"airtime": "MON-FRI 9pm-10pm, TUE-FRI 1am-2am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/BBC-World-Service-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.bbc.co.uk/sounds/play/live:bbc_world_service",
"meta": {
"site": "news",
"source": "BBC World Service"
},
"link": "/radio/program/bbc-world-service",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/global-news-podcast/id135067274?mt=2",
"tuneIn": "https://tunein.com/radio/BBC-World-Service-p455581/",
"rss": "https://podcasts.files.bbci.co.uk/p02nq0gn.rss"
}
},
"californiareport": {
"id": "californiareport",
"title": "The California Report",
"tagline": "California, day by day",
"info": "KQED’s statewide radio news program providing daily coverage of issues, trends and public policy decisions.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-California-Report-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The California Report",
"officialWebsiteLink": "/californiareport",
"meta": {
"site": "news",
"source": "kqed",
"order": 8
},
"link": "/californiareport",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/kqeds-the-california-report/id79681292",
"amazon": "https://music.amazon.com/podcasts/26099305-72af-4542-9dde-ac1807fe36d5/kqed-s-the-california-report",
"npr": "https://www.npr.org/podcasts/432285393/the-california-report",
"stitcher": "https://www.stitcher.com/podcast/kqedfm-kqeds-the-california-report-podcast-8838",
"rss": "https://ww2.kqed.org/news/tag/tcram/feed/podcast"
}
},
"californiareportmagazine": {
"id": "californiareportmagazine",
"title": "The California Report Magazine",
"tagline": "Your state, your stories",
"info": "Every week, The California Report Magazine takes you on a road trip for the ears: to visit the places and meet the people who make California unique. The in-depth storytelling podcast from the California Report.",
"airtime": "FRI 4:30pm-5pm, 6:30pm-7pm, 11pm-11:30pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-California-Report-Magazine-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The California Report Magazine",
"officialWebsiteLink": "/californiareportmagazine",
"meta": {
"site": "news",
"source": "kqed",
"order": 10
},
"link": "/californiareportmagazine",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/the-california-report-magazine/id1314750545",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vS1FJTkM3NjkwNjk1OTAz",
"npr": "https://www.npr.org/podcasts/564733126/the-california-report-magazine",
"stitcher": "https://www.stitcher.com/podcast/kqed/the-california-report-magazine",
"rss": "https://ww2.kqed.org/news/tag/tcrmag/feed/podcast"
}
},
"city-arts": {
"id": "city-arts",
"title": "City Arts & Lectures",
"info": "A one-hour radio program to hear celebrated writers, artists and thinkers address contemporary ideas and values, often discussing the creative process. Please note: tapes or transcripts are not available",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/05/cityartsandlecture-300x300.jpg",
"officialWebsiteLink": "https://www.cityarts.net/",
"airtime": "SUN 1pm-2pm, TUE 10pm, WED 1am",
"meta": {
"site": "news",
"source": "City Arts & Lectures"
},
"link": "https://www.cityarts.net",
"subscribe": {
"tuneIn": "https://tunein.com/radio/City-Arts-and-Lectures-p692/",
"rss": "https://www.cityarts.net/feed/"
}
},
"closealltabs": {
"id": "closealltabs",
"title": "Close All Tabs",
"tagline": "Your irreverent guide to the trends redefining our world",
"info": "Close All Tabs breaks down how digital culture shapes our world through thoughtful insights and irreverent humor.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/02/CAT_2_Tile-scaled.jpg",
"imageAlt": "KQED Close All Tabs",
"officialWebsiteLink": "/podcasts/closealltabs",
"meta": {
"site": "news",
"source": "kqed",
"order": 1
},
"link": "/podcasts/closealltabs",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/close-all-tabs/id214663465",
"rss": "https://feeds.megaphone.fm/KQINC6993880386",
"amazon": "https://music.amazon.com/podcasts/92d9d4ac-67a3-4eed-b10a-fb45d45b1ef2/close-all-tabs",
"spotify": "https://open.spotify.com/show/6LAJFHnGK1pYXYzv6SIol6?si=deb0cae19813417c"
}
},
"code-switch-life-kit": {
"id": "code-switch-life-kit",
"title": "Code Switch / Life Kit",
"info": "\u003cem>Code Switch\u003c/em>, which listeners will hear in the first part of the hour, has fearless and much-needed conversations about race. Hosted by journalists of color, the show tackles the subject of race head-on, exploring how it impacts every part of society — from politics and pop culture to history, sports and more.\u003cbr />\u003cbr />\u003cem>Life Kit\u003c/em>, which will be in the second part of the hour, guides you through spaces and feelings no one prepares you for — from finances to mental health, from workplace microaggressions to imposter syndrome, from relationships to parenting. The show features experts with real world experience and shares their knowledge. Because everyone needs a little help being human.\u003cbr />\u003cbr />\u003ca href=\"https://www.npr.org/podcasts/510312/codeswitch\">\u003cem>Code Switch\u003c/em> offical site and podcast\u003c/a>\u003cbr />\u003ca href=\"https://www.npr.org/lifekit\">\u003cem>Life Kit\u003c/em> offical site and podcast\u003c/a>\u003cbr />",
"airtime": "SUN 9pm-10pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Code-Switch-Life-Kit-Podcast-Tile-360x360-1.jpg",
"meta": {
"site": "radio",
"source": "npr"
},
"link": "/radio/program/code-switch-life-kit",
"subscribe": {
"apple": "https://podcasts.apple.com/podcast/1112190608?mt=2&at=11l79Y&ct=nprdirectory",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly93d3cubnByLm9yZy9yc3MvcG9kY2FzdC5waHA_aWQ9NTEwMzEy",
"spotify": "https://open.spotify.com/show/3bExJ9JQpkwNhoHvaIIuyV",
"rss": "https://feeds.npr.org/510312/podcast.xml"
}
},
"commonwealth-club": {
"id": "commonwealth-club",
"title": "Commonwealth Club of California Podcast",
"info": "The Commonwealth Club of California is the nation's oldest and largest public affairs forum. As a non-partisan forum, The Club brings to the public airwaves diverse viewpoints on important topics. The Club's weekly radio broadcast - the oldest in the U.S., dating back to 1924 - is carried across the nation on public radio stations and is now podcasting. Our website archive features audio of our recent programs, as well as selected speeches from our long and distinguished history. This podcast feed is usually updated twice a week and is always un-edited.",
"airtime": "THU 10pm, FRI 1am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Commonwealth-Club-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.commonwealthclub.org/podcasts",
"meta": {
"site": "news",
"source": "Commonwealth Club of California"
},
"link": "/radio/program/commonwealth-club",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/commonwealth-club-of-california-podcast/id976334034?mt=2",
"google": "https://podcasts.google.com/feed/aHR0cDovL3d3dy5jb21tb253ZWFsdGhjbHViLm9yZy9hdWRpby9wb2RjYXN0L3dlZWtseS54bWw",
"tuneIn": "https://tunein.com/radio/Commonwealth-Club-of-California-p1060/"
}
},
"forum": {
"id": "forum",
"title": "Forum",
"tagline": "The conversation starts here",
"info": "KQED’s live call-in program discussing local, state, national and international issues, as well as in-depth interviews.",
"airtime": "MON-FRI 9am-11am, 10pm-11pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Forum-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Forum with Mina Kim and Alexis Madrigal",
"officialWebsiteLink": "/forum",
"meta": {
"site": "news",
"source": "kqed",
"order": 9
},
"link": "/forum",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/kqeds-forum/id73329719",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vS1FJTkM5NTU3MzgxNjMz",
"npr": "https://www.npr.org/podcasts/432307980/forum",
"stitcher": "https://www.stitcher.com/podcast/kqedfm-kqeds-forum-podcast",
"rss": "https://feeds.megaphone.fm/KQINC9557381633"
}
},
"freakonomics-radio": {
"id": "freakonomics-radio",
"title": "Freakonomics Radio",
"info": "Freakonomics Radio is a one-hour award-winning podcast and public-radio project hosted by Stephen Dubner, with co-author Steve Levitt as a regular guest. It is produced in partnership with WNYC.",
"imageSrc": "https://ww2.kqed.org/news/wp-content/uploads/sites/10/2018/05/freakonomicsRadio.png",
"officialWebsiteLink": "http://freakonomics.com/",
"airtime": "SUN 1am-2am, SAT 3pm-4pm",
"meta": {
"site": "radio",
"source": "WNYC"
},
"link": "/radio/program/freakonomics-radio",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/4s8b",
"apple": "https://itunes.apple.com/us/podcast/freakonomics-radio/id354668519",
"tuneIn": "https://tunein.com/podcasts/WNYC-Podcasts/Freakonomics-Radio-p272293/",
"rss": "https://feeds.feedburner.com/freakonomicsradio"
}
},
"fresh-air": {
"id": "fresh-air",
"title": "Fresh Air",
"info": "Hosted by Terry Gross, \u003cem>Fresh Air from WHYY\u003c/em> is the Peabody Award-winning weekday magazine of contemporary arts and issues. One of public radio's most popular programs, Fresh Air features intimate conversations with today's biggest luminaries.",
"airtime": "MON-FRI 7pm-8pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Fresh-Air-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/fresh-air/",
"meta": {
"site": "radio",
"source": "npr"
},
"link": "/radio/program/fresh-air",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/4s8b",
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=214089682&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/Fresh-Air-p17/",
"rss": "https://feeds.npr.org/381444908/podcast.xml"
}
},
"here-and-now": {
"id": "here-and-now",
"title": "Here & Now",
"info": "A live production of NPR and WBUR Boston, in collaboration with stations across the country, Here & Now reflects the fluid world of news as it's happening in the middle of the day, with timely, in-depth news, interviews and conversation. Hosted by Robin Young, Jeremy Hobson and Tonya Mosley.",
"airtime": "MON-THU 11am-12pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Here-And-Now-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "http://www.wbur.org/hereandnow",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/here-and-now",
"subsdcribe": {
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?mt=2&id=426698661",
"tuneIn": "https://tunein.com/radio/Here--Now-p211/",
"rss": "https://feeds.npr.org/510051/podcast.xml"
}
},
"hidden-brain": {
"id": "hidden-brain",
"title": "Hidden Brain",
"info": "Shankar Vedantam uses science and storytelling to reveal the unconscious patterns that drive human behavior, shape our choices and direct our relationships.",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/05/hiddenbrain.jpg",
"officialWebsiteLink": "https://www.npr.org/series/423302056/hidden-brain",
"airtime": "SUN 7pm-8pm",
"meta": {
"site": "news",
"source": "NPR"
},
"link": "/radio/program/hidden-brain",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/hidden-brain/id1028908750?mt=2",
"tuneIn": "https://tunein.com/podcasts/Science-Podcasts/Hidden-Brain-p787503/",
"rss": "https://feeds.npr.org/510308/podcast.xml"
}
},
"how-i-built-this": {
"id": "how-i-built-this",
"title": "How I Built This with Guy Raz",
"info": "Guy Raz dives into the stories behind some of the world's best known companies. How I Built This weaves a narrative journey about innovators, entrepreneurs and idealists—and the movements they built.",
"imageSrc": "https://ww2.kqed.org/news/wp-content/uploads/sites/10/2018/05/howIBuiltThis.png",
"officialWebsiteLink": "https://www.npr.org/podcasts/510313/how-i-built-this",
"airtime": "SUN 7:30pm-8pm",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/how-i-built-this",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/3zxy",
"apple": "https://itunes.apple.com/us/podcast/how-i-built-this-with-guy-raz/id1150510297?mt=2",
"tuneIn": "https://tunein.com/podcasts/Arts--Culture-Podcasts/How-I-Built-This-p910896/",
"rss": "https://feeds.npr.org/510313/podcast.xml"
}
},
"hyphenacion": {
"id": "hyphenacion",
"title": "Hyphenación",
"tagline": "Where conversation and cultura meet",
"info": "What kind of no sabo word is Hyphenación? For us, it’s about living within a hyphenation. Like being a third-gen Mexican-American from the Texas border now living that Bay Area Chicano life. Like Xorje! Each week we bring together a couple of hyphenated Latinos to talk all about personal life choices: family, careers, relationships, belonging … everything is on the table. ",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/03/Hyphenacion_FinalAssets_PodcastTile.png",
"imageAlt": "KQED Hyphenación",
"officialWebsiteLink": "/podcasts/hyphenacion",
"meta": {
"site": "news",
"source": "kqed",
"order": 15
},
"link": "/podcasts/hyphenacion",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/hyphenaci%C3%B3n/id1191591838",
"spotify": "https://open.spotify.com/show/2p3Fifq96nw9BPcmFdIq0o?si=39209f7b25774f38",
"youtube": "https://www.youtube.com/c/kqedarts",
"amazon": "https://music.amazon.com/podcasts/6c3dd23c-93fb-4aab-97ba-1725fa6315f1/hyphenaci%C3%B3n",
"rss": "https://feeds.megaphone.fm/KQINC2275451163"
}
},
"jerrybrown": {
"id": "jerrybrown",
"title": "The Political Mind of Jerry Brown",
"tagline": "Lessons from a lifetime in politics",
"info": "The Political Mind of Jerry Brown brings listeners the wisdom of the former Governor, Mayor, and presidential candidate. Scott Shafer interviewed Brown for more than 40 hours, covering the former governor's life and half-century in the political game – and Brown has some lessons he'd like to share. ",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-Political-Mind-of-Jerry-Brown-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The Political Mind of Jerry Brown",
"officialWebsiteLink": "/podcasts/jerrybrown",
"meta": {
"site": "news",
"source": "kqed",
"order": 18
},
"link": "/podcasts/jerrybrown",
"subscribe": {
"npr": "https://www.npr.org/podcasts/790253322/the-political-mind-of-jerry-brown",
"apple": "https://itunes.apple.com/us/podcast/id1492194549",
"rss": "https://ww2.kqed.org/news/series/jerrybrown/feed/podcast/",
"tuneIn": "http://tun.in/pjGcK",
"stitcher": "https://www.stitcher.com/podcast/kqed/the-political-mind-of-jerry-brown",
"spotify": "https://open.spotify.com/show/54C1dmuyFyKMFttY6X2j6r?si=K8SgRCoISNK6ZbjpXrX5-w",
"amazon": "https://music.amazon.com/podcasts/44420f75-3b0e-4301-ab3b-16da6b09e543/the-political-mind-of-jerry-brown"
}
},
"latino-usa": {
"id": "latino-usa",
"title": "Latino USA",
"airtime": "MON 1am-2am, SUN 6pm-7pm",
"info": "Latino USA, the radio journal of news and culture, is the only national, English-language radio program produced from a Latino perspective.",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/latinoUsa.jpg",
"officialWebsiteLink": "http://latinousa.org/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/latino-usa",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/xtTd",
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=79681317&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/Latino-USA-p621/",
"rss": "https://feeds.npr.org/510016/podcast.xml"
}
},
"marketplace": {
"id": "marketplace",
"title": "Marketplace",
"info": "Our flagship program, helmed by Kai Ryssdal, examines what the day in money delivered, through stories, conversations, newsworthy numbers and more. Updated Monday through Friday at about 3:30 p.m. PT.",
"airtime": "MON-FRI 4pm-4:30pm, MON-WED 6:30pm-7pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Marketplace-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.marketplace.org/",
"meta": {
"site": "news",
"source": "American Public Media"
},
"link": "/radio/program/marketplace",
"subscribe": {
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=201853034&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/APM-Marketplace-p88/",
"rss": "https://feeds.publicradio.org/public_feeds/marketplace-pm/rss/rss"
}
},
"masters-of-scale": {
"id": "masters-of-scale",
"title": "Masters of Scale",
"info": "Masters of Scale is an original podcast in which LinkedIn co-founder and Greylock Partner Reid Hoffman sets out to describe and prove theories that explain how great entrepreneurs take their companies from zero to a gazillion in ingenious fashion.",
"airtime": "Every other Wednesday June 12 through October 16 at 8pm (repeats Thursdays at 2am)",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Masters-of-Scale-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://mastersofscale.com/",
"meta": {
"site": "radio",
"source": "WaitWhat"
},
"link": "/radio/program/masters-of-scale",
"subscribe": {
"apple": "http://mastersofscale.app.link/",
"rss": "https://rss.art19.com/masters-of-scale"
}
},
"mindshift": {
"id": "mindshift",
"title": "MindShift",
"tagline": "A podcast about the future of learning and how we raise our kids",
"info": "The MindShift podcast explores the innovations in education that are shaping how kids learn. Hosts Ki Sung and Katrina Schwartz introduce listeners to educators, researchers, parents and students who are developing effective ways to improve how kids learn. We cover topics like how fed-up administrators are developing surprising tactics to deal with classroom disruptions; how listening to podcasts are helping kids develop reading skills; the consequences of overparenting; and why interdisciplinary learning can engage students on all ends of the traditional achievement spectrum. This podcast is part of the MindShift education site, a division of KQED News. KQED is an NPR/PBS member station based in San Francisco. You can also visit the MindShift website for episodes and supplemental blog posts or tweet us \u003ca href=\"https://twitter.com/MindShiftKQED\">@MindShiftKQED\u003c/a> or visit us at \u003ca href=\"/mindshift\">MindShift.KQED.org\u003c/a>",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Mindshift-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED MindShift: How We Will Learn",
"officialWebsiteLink": "/mindshift/",
"meta": {
"site": "news",
"source": "kqed",
"order": 12
},
"link": "/podcasts/mindshift",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/mindshift-podcast/id1078765985",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vS1FJTkM1NzY0NjAwNDI5",
"npr": "https://www.npr.org/podcasts/464615685/mind-shift-podcast",
"stitcher": "https://www.stitcher.com/podcast/kqed/stories-teachers-share",
"spotify": "https://open.spotify.com/show/0MxSpNYZKNprFLCl7eEtyx"
}
},
"morning-edition": {
"id": "morning-edition",
"title": "Morning Edition",
"info": "\u003cem>Morning Edition\u003c/em> takes listeners around the country and the world with multi-faceted stories and commentaries every weekday. Hosts Steve Inskeep, David Greene and Rachel Martin bring you the latest breaking news and features to prepare you for the day.",
"airtime": "MON-FRI 3am-9am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Morning-Edition-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/morning-edition/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/morning-edition"
},
"onourwatch": {
"id": "onourwatch",
"title": "On Our Watch",
"tagline": "Deeply-reported investigative journalism",
"info": "For decades, the process for how police police themselves has been inconsistent – if not opaque. In some states, like California, these proceedings were completely hidden. After a new police transparency law unsealed scores of internal affairs files, our reporters set out to examine these cases and the shadow world of police discipline. On Our Watch brings listeners into the rooms where officers are questioned and witnesses are interrogated to find out who this system is really protecting. Is it the officers, or the public they've sworn to serve?",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/On-Our-Watch-Podcast-Tile-703x703-1.jpg",
"imageAlt": "On Our Watch from NPR and KQED",
"officialWebsiteLink": "/podcasts/onourwatch",
"meta": {
"site": "news",
"source": "kqed",
"order": 11
},
"link": "/podcasts/onourwatch",
"subscribe": {
"apple": "https://podcasts.apple.com/podcast/id1567098962",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5ucHIub3JnLzUxMDM2MC9wb2RjYXN0LnhtbD9zYz1nb29nbGVwb2RjYXN0cw",
"npr": "https://rpb3r.app.goo.gl/onourwatch",
"spotify": "https://open.spotify.com/show/0OLWoyizopu6tY1XiuX70x",
"tuneIn": "https://tunein.com/radio/On-Our-Watch-p1436229/",
"stitcher": "https://www.stitcher.com/show/on-our-watch",
"rss": "https://feeds.npr.org/510360/podcast.xml"
}
},
"on-the-media": {
"id": "on-the-media",
"title": "On The Media",
"info": "Our weekly podcast explores how the media 'sausage' is made, casts an incisive eye on fluctuations in the marketplace of ideas, and examines threats to the freedom of information and expression in America and abroad. For one hour a week, the show tries to lift the veil from the process of \"making media,\" especially news media, because it's through that lens that we see the world and the world sees us",
"airtime": "SUN 2pm-3pm, MON 12am-1am",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/onTheMedia.png",
"officialWebsiteLink": "https://www.wnycstudios.org/shows/otm",
"meta": {
"site": "news",
"source": "wnyc"
},
"link": "/radio/program/on-the-media",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/on-the-media/id73330715?mt=2",
"tuneIn": "https://tunein.com/radio/On-the-Media-p69/",
"rss": "http://feeds.wnyc.org/onthemedia"
}
},
"pbs-newshour": {
"id": "pbs-newshour",
"title": "PBS NewsHour",
"info": "Analysis, background reports and updates from the PBS NewsHour putting today's news in context.",
"airtime": "MON-FRI 3pm-4pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/PBS-News-Hour-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.pbs.org/newshour/",
"meta": {
"site": "news",
"source": "pbs"
},
"link": "/radio/program/pbs-newshour",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/pbs-newshour-full-show/id394432287?mt=2",
"tuneIn": "https://tunein.com/radio/PBS-NewsHour---Full-Show-p425698/",
"rss": "https://www.pbs.org/newshour/feeds/rss/podcasts/show"
}
},
"perspectives": {
"id": "perspectives",
"title": "Perspectives",
"tagline": "KQED's series of daily listener commentaries since 1991",
"info": "KQED's series of daily listener commentaries since 1991.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/01/Perspectives_Tile_Final.jpg",
"imageAlt": "KQED Perspectives",
"officialWebsiteLink": "/perspectives/",
"meta": {
"site": "radio",
"source": "kqed",
"order": 14
},
"link": "/perspectives",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/id73801135",
"npr": "https://www.npr.org/podcasts/432309616/perspectives",
"rss": "https://ww2.kqed.org/perspectives/category/perspectives/feed/",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly93dzIua3FlZC5vcmcvcGVyc3BlY3RpdmVzL2NhdGVnb3J5L3BlcnNwZWN0aXZlcy9mZWVkLw"
}
},
"planet-money": {
"id": "planet-money",
"title": "Planet Money",
"info": "The economy explained. Imagine you could call up a friend and say, Meet me at the bar and tell me what's going on with the economy. Now imagine that's actually a fun evening.",
"airtime": "SUN 3pm-4pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/planetmoney.jpg",
"officialWebsiteLink": "https://www.npr.org/sections/money/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/planet-money",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/M4f5",
"apple": "https://itunes.apple.com/us/podcast/planet-money/id290783428?mt=2",
"tuneIn": "https://tunein.com/podcasts/Business--Economics-Podcasts/Planet-Money-p164680/",
"rss": "https://feeds.npr.org/510289/podcast.xml"
}
},
"politicalbreakdown": {
"id": "politicalbreakdown",
"title": "Political Breakdown",
"tagline": "Politics from a personal perspective",
"info": "Political Breakdown is a new series that explores the political intersection of California and the nation. Each week hosts Scott Shafer and Marisa Lagos are joined with a new special guest to unpack politics -- with personality — and offer an insider’s glimpse at how politics happens.",
"airtime": "THU 6:30pm-7pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Political-Breakdown-2024-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Political Breakdown",
"officialWebsiteLink": "/podcasts/politicalbreakdown",
"meta": {
"site": "radio",
"source": "kqed",
"order": 5
},
"link": "/podcasts/politicalbreakdown",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/political-breakdown/id1327641087",
"amazon": "https://music.amazon.com/podcasts/e0c2d153-ad36-4c8d-901d-f1da6a724824/political-breakdown",
"npr": "https://www.npr.org/podcasts/572155894/political-breakdown",
"stitcher": "https://www.stitcher.com/podcast/kqed/political-breakdown",
"spotify": "https://open.spotify.com/show/07RVyIjIdk2WDuVehvBMoN",
"rss": "https://ww2.kqed.org/news/tag/political-breakdown/feed/podcast"
}
},
"possible": {
"id": "possible",
"title": "Possible",
"info": "Possible is hosted by entrepreneur Reid Hoffman and writer Aria Finger. Together in Possible, Hoffman and Finger lead enlightening discussions about building a brighter collective future. The show features interviews with visionary guests like Trevor Noah, Sam Altman and Janette Sadik-Khan. Possible paints an optimistic portrait of the world we can create through science, policy, business, art and our shared humanity. It asks: What if everything goes right for once? How can we get there? Each episode also includes a short fiction story generated by advanced AI GPT-4, serving as a thought-provoking springboard to speculate how humanity could leverage technology for good.",
"airtime": "SUN 2pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Possible-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.possible.fm/",
"meta": {
"site": "news",
"source": "Possible"
},
"link": "/radio/program/possible",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/possible/id1677184070",
"spotify": "https://open.spotify.com/show/730YpdUSNlMyPQwNnyjp4k"
}
},
"pri-the-world": {
"id": "pri-the-world",
"title": "PRI's The World: Latest Edition",
"info": "Each weekday, host Marco Werman and his team of producers bring you the world's most interesting stories in an hour of radio that reminds us just how small our planet really is.",
"airtime": "MON-FRI 2pm-3pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-World-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.pri.org/programs/the-world",
"meta": {
"site": "news",
"source": "PRI"
},
"link": "/radio/program/pri-the-world",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/pris-the-world-latest-edition/id278196007?mt=2",
"tuneIn": "https://tunein.com/podcasts/News--Politics-Podcasts/PRIs-The-World-p24/",
"rss": "http://feeds.feedburner.com/pri/theworld"
}
},
"radiolab": {
"id": "radiolab",
"title": "Radiolab",
"info": "A two-time Peabody Award-winner, Radiolab is an investigation told through sounds and stories, and centered around one big idea. In the Radiolab world, information sounds like music and science and culture collide. Hosted by Jad Abumrad and Robert Krulwich, the show is designed for listeners who demand skepticism, but appreciate wonder. WNYC Studios is the producer of other leading podcasts including Freakonomics Radio, Death, Sex & Money, On the Media and many more.",
"airtime": "SUN 12am-1am, SAT 2pm-3pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/radiolab1400.png",
"officialWebsiteLink": "https://www.wnycstudios.org/shows/radiolab/",
"meta": {
"site": "science",
"source": "WNYC"
},
"link": "/radio/program/radiolab",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/radiolab/id152249110?mt=2",
"tuneIn": "https://tunein.com/radio/RadioLab-p68032/",
"rss": "https://feeds.wnyc.org/radiolab"
}
},
"reveal": {
"id": "reveal",
"title": "Reveal",
"info": "Created by The Center for Investigative Reporting and PRX, Reveal is public radios first one-hour weekly radio show and podcast dedicated to investigative reporting. Credible, fact based and without a partisan agenda, Reveal combines the power and artistry of driveway moment storytelling with data-rich reporting on critically important issues. The result is stories that inform and inspire, arming our listeners with information to right injustices, hold the powerful accountable and improve lives.Reveal is hosted by Al Letson and showcases the award-winning work of CIR and newsrooms large and small across the nation. In a radio and podcast market crowded with choices, Reveal focuses on important and often surprising stories that illuminate the world for our listeners.",
"airtime": "SAT 4pm-5pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/reveal300px.png",
"officialWebsiteLink": "https://www.revealnews.org/episodes/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/reveal",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/reveal/id886009669",
"tuneIn": "https://tunein.com/radio/Reveal-p679597/",
"rss": "http://feeds.revealradio.org/revealpodcast"
}
},
"rightnowish": {
"id": "rightnowish",
"title": "Rightnowish",
"tagline": "Art is where you find it",
"info": "Rightnowish digs into life in the Bay Area right now… ish. Journalist Pendarvis Harshaw takes us to galleries painted on the sides of liquor stores in West Oakland. We'll dance in warehouses in the Bayview, make smoothies with kids in South Berkeley, and listen to classical music in a 1984 Cutlass Supreme in Richmond. Every week, Pen talks to movers and shakers about how the Bay Area shapes what they create, and how they shape the place we call home.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Rightnowish-Podcast-Tile-500x500-1.jpg",
"imageAlt": "KQED Rightnowish with Pendarvis Harshaw",
"officialWebsiteLink": "/podcasts/rightnowish",
"meta": {
"site": "arts",
"source": "kqed",
"order": 16
},
"link": "/podcasts/rightnowish",
"subscribe": {
"npr": "https://www.npr.org/podcasts/721590300/rightnowish",
"rss": "https://ww2.kqed.org/arts/programs/rightnowish/feed/podcast",
"apple": "https://podcasts.apple.com/us/podcast/rightnowish/id1482187648",
"stitcher": "https://www.stitcher.com/podcast/kqed/rightnowish",
"google": "https://podcasts.google.com/feed/aHR0cHM6Ly9mZWVkcy5tZWdhcGhvbmUuZm0vS1FJTkMxMjU5MTY3NDc4",
"spotify": "https://open.spotify.com/show/7kEJuafTzTVan7B78ttz1I"
}
},
"science-friday": {
"id": "science-friday",
"title": "Science Friday",
"info": "Science Friday is a weekly science talk show, broadcast live over public radio stations nationwide. Each week, the show focuses on science topics that are in the news and tries to bring an educated, balanced discussion to bear on the scientific issues at hand. Panels of expert guests join host Ira Flatow, a veteran science journalist, to discuss science and to take questions from listeners during the call-in portion of the program.",
"airtime": "FRI 11am-1pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Science-Friday-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.wnycstudios.org/shows/science-friday",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/science-friday",
"subscribe": {
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=73329284&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/Science-Friday-p394/",
"rss": "http://feeds.wnyc.org/science-friday"
}
},
"snap-judgment": {
"id": "snap-judgment",
"title": "Snap Judgment",
"tagline": "Real stories with killer beats",
"info": "The Snap Judgment radio show and podcast mixes real stories with killer beats to produce cinematic, dramatic radio. Snap's musical brand of storytelling dares listeners to see the world through the eyes of another. This is storytelling... with a BEAT!! Snap first aired on public radio stations nationwide in July 2010. Today, Snap Judgment airs on over 450 public radio stations and is brought to the airwaves by KQED & PRX.",
"airtime": "SAT 1pm-2pm, 9pm-10pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/05/Snap-Judgment-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Snap Judgment",
"officialWebsiteLink": "https://snapjudgment.org",
"meta": {
"site": "arts",
"source": "kqed",
"order": 4
},
"link": "https://snapjudgment.org",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/snap-judgment/id283657561",
"npr": "https://www.npr.org/podcasts/449018144/snap-judgment",
"stitcher": "https://www.pandora.com/podcast/snap-judgment/PC:241?source=stitcher-sunset",
"spotify": "https://open.spotify.com/show/3Cct7ZWmxHNAtLgBTqjC5v",
"rss": "https://snap.feed.snapjudgment.org/"
}
},
"soldout": {
"id": "soldout",
"title": "SOLD OUT: Rethinking Housing in America",
"tagline": "A new future for housing",
"info": "Sold Out: Rethinking Housing in America",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Sold-Out-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Sold Out: Rethinking Housing in America",
"officialWebsiteLink": "/podcasts/soldout",
"meta": {
"site": "news",
"source": "kqed",
"order": 13
},
"link": "/podcasts/soldout",
"subscribe": {
"npr": "https://www.npr.org/podcasts/911586047/s-o-l-d-o-u-t-a-new-future-for-housing",
"apple": "https://podcasts.apple.com/us/podcast/introducing-sold-out-rethinking-housing-in-america/id1531354937",
"rss": "https://feeds.megaphone.fm/soldout",
"spotify": "https://open.spotify.com/show/38dTBSk2ISFoPiyYNoKn1X",
"stitcher": "https://www.stitcher.com/podcast/kqed/sold-out-rethinking-housing-in-america",
"tunein": "https://tunein.com/radio/SOLD-OUT-Rethinking-Housing-in-America-p1365871/"
}
},
"spooked": {
"id": "spooked",
"title": "Spooked",
"tagline": "True-life supernatural stories",
"info": "",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/10/Spooked-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED Spooked",
"officialWebsiteLink": "https://spookedpodcast.org/",
"meta": {
"site": "news",
"source": "kqed",
"order": 7
},
"link": "https://spookedpodcast.org/",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/spooked/id1279361017",
"npr": "https://www.npr.org/podcasts/549547848/snap-judgment-presents-spooked",
"spotify": "https://open.spotify.com/show/76571Rfl3m7PLJQZKQIGCT",
"rss": "https://feeds.simplecast.com/TBotaapn"
}
},
"tech-nation": {
"id": "tech-nation",
"title": "Tech Nation Radio Podcast",
"info": "Tech Nation is a weekly public radio program, hosted by Dr. Moira Gunn. Founded in 1993, it has grown from a simple interview show to a multi-faceted production, featuring conversations with noted technology and science leaders, and a weekly science and technology-related commentary.",
"airtime": "FRI 10pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Tech-Nation-Radio-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "http://technation.podomatic.com/",
"meta": {
"site": "science",
"source": "Tech Nation Media"
},
"link": "/radio/program/tech-nation",
"subscribe": {
"rss": "https://technation.podomatic.com/rss2.xml"
}
},
"ted-radio-hour": {
"id": "ted-radio-hour",
"title": "TED Radio Hour",
"info": "The TED Radio Hour is a journey through fascinating ideas, astonishing inventions, fresh approaches to old problems, and new ways to think and create.",
"airtime": "SUN 3pm-4pm, SAT 10pm-11pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/tedRadioHour.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/ted-radio-hour/?showDate=2018-06-22",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/ted-radio-hour",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/8vsS",
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=523121474&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/TED-Radio-Hour-p418021/",
"rss": "https://feeds.npr.org/510298/podcast.xml"
}
},
"thebay": {
"id": "thebay",
"title": "The Bay",
"tagline": "Local news to keep you rooted",
"info": "Host Devin Katayama walks you through the biggest story of the day with reporters and newsmakers.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-Bay-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The Bay",
"officialWebsiteLink": "/podcasts/thebay",
"meta": {
"site": "radio",
"source": "kqed",
"order": 2
},
"link": "/podcasts/thebay",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/the-bay/id1350043452",
"amazon": "https://music.amazon.com/podcasts/d800ea4c-7a2c-42f2-b861-edaf78a5db0b/the-bay",
"npr": "https://www.npr.org/podcasts/586725995/the-bay",
"stitcher": "https://www.stitcher.com/podcast/kqed/the-bay",
"spotify": "https://open.spotify.com/show/4BIKBKIujizLHlIlBNaAqQ",
"rss": "https://feeds.megaphone.fm/KQINC8259786327"
}
},
"thelatest": {
"id": "thelatest",
"title": "The Latest",
"tagline": "Trusted local news in real time",
"info": "",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/05/The-Latest-2025-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The Latest",
"officialWebsiteLink": "/thelatest",
"meta": {
"site": "news",
"source": "kqed",
"order": 6
},
"link": "/thelatest",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/the-latest-from-kqed/id1197721799",
"npr": "https://www.npr.org/podcasts/1257949365/the-latest-from-k-q-e-d",
"spotify": "https://open.spotify.com/show/5KIIXMgM9GTi5AepwOYvIZ?si=bd3053fec7244dba",
"rss": "https://feeds.megaphone.fm/KQINC9137121918"
}
},
"theleap": {
"id": "theleap",
"title": "The Leap",
"tagline": "What if you closed your eyes, and jumped?",
"info": "Stories about people making dramatic, risky changes, told by award-winning public radio reporter Judy Campbell.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-Leap-Podcast-Tile-703x703-1.jpg",
"imageAlt": "KQED The Leap",
"officialWebsiteLink": "/podcasts/theleap",
"meta": {
"site": "news",
"source": "kqed",
"order": 17
},
"link": "/podcasts/theleap",
"subscribe": {
"apple": "https://podcasts.apple.com/us/podcast/the-leap/id1046668171",
"npr": "https://www.npr.org/podcasts/447248267/the-leap",
"stitcher": "https://www.stitcher.com/podcast/kqed/the-leap",
"spotify": "https://open.spotify.com/show/3sSlVHHzU0ytLwuGs1SD1U",
"rss": "https://ww2.kqed.org/news/programs/the-leap/feed/podcast"
}
},
"the-moth-radio-hour": {
"id": "the-moth-radio-hour",
"title": "The Moth Radio Hour",
"info": "Since its launch in 1997, The Moth has presented thousands of true stories, told live and without notes, to standing-room-only crowds worldwide. Moth storytellers stand alone, under a spotlight, with only a microphone and a roomful of strangers. The storyteller and the audience embark on a high-wire act of shared experience which is both terrifying and exhilarating. Since 2008, The Moth podcast has featured many of our favorite stories told live on Moth stages around the country. For information on all of our programs and live events, visit themoth.org.",
"airtime": "SAT 8pm-9pm and SUN 11am-12pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/theMoth.jpg",
"officialWebsiteLink": "https://themoth.org/",
"meta": {
"site": "arts",
"source": "prx"
},
"link": "/radio/program/the-moth-radio-hour",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/the-moth-podcast/id275699983?mt=2",
"tuneIn": "https://tunein.com/radio/The-Moth-p273888/",
"rss": "http://feeds.themoth.org/themothpodcast"
}
},
"the-new-yorker-radio-hour": {
"id": "the-new-yorker-radio-hour",
"title": "The New Yorker Radio Hour",
"info": "The New Yorker Radio Hour is a weekly program presented by the magazine's editor, David Remnick, and produced by WNYC Studios and The New Yorker. Each episode features a diverse mix of interviews, profiles, storytelling, and an occasional burst of humor inspired by the magazine, and shaped by its writers, artists, and editors. This isn't a radio version of a magazine, but something all its own, reflecting the rich possibilities of audio storytelling and conversation. Theme music for the show was composed and performed by Merrill Garbus of tUnE-YArDs.",
"airtime": "SAT 10am-11am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-New-Yorker-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.wnycstudios.org/shows/tnyradiohour",
"meta": {
"site": "arts",
"source": "WNYC"
},
"link": "/radio/program/the-new-yorker-radio-hour",
"subscribe": {
"apple": "https://itunes.apple.com/us/podcast/id1050430296",
"tuneIn": "https://tunein.com/podcasts/WNYC-Podcasts/New-Yorker-Radio-Hour-p803804/",
"rss": "https://feeds.feedburner.com/newyorkerradiohour"
}
},
"the-sam-sanders-show": {
"id": "the-sam-sanders-show",
"title": "The Sam Sanders Show",
"info": "One of public radio's most dynamic voices, Sam Sanders helped launch The NPR Politics Podcast and hosted NPR's hit show It's Been A Minute. Now, the award-winning host returns with something brand new, The Sam Sanders Show. Every week, Sam Sanders and friends dig into the culture that shapes our lives: what's driving the biggest trends, how artists really think, and even the memes you can't stop scrolling past. Sam is beloved for his way of unpacking the world and bringing you up close to fresh currents and engaging conversations. The Sam Sanders Show is smart, funny and always a good time.",
"airtime": "FRI 12-1pm AND SAT 11am-12pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/11/The-Sam-Sanders-Show-Podcast-Tile-400x400-1.jpg",
"officialWebsiteLink": "https://www.kcrw.com/shows/the-sam-sanders-show/latest",
"meta": {
"site": "arts",
"source": "KCRW"
},
"link": "https://www.kcrw.com/shows/the-sam-sanders-show/latest",
"subscribe": {
"rss": "https://feed.cdnstream1.com/zjb/feed/download/ac/28/59/ac28594c-e1d0-4231-8728-61865cdc80e8.xml"
}
},
"the-splendid-table": {
"id": "the-splendid-table",
"title": "The Splendid Table",
"info": "\u003cem>The Splendid Table\u003c/em> hosts our nation's conversations about cooking, sustainability and food culture.",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/The-Splendid-Table-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.splendidtable.org/",
"airtime": "SUN 10-11 pm",
"meta": {
"site": "radio",
"source": "npr"
},
"link": "/radio/program/the-splendid-table"
},
"this-american-life": {
"id": "this-american-life",
"title": "This American Life",
"info": "This American Life is a weekly public radio show, heard by 2.2 million people on more than 500 stations. Another 2.5 million people download the weekly podcast. It is hosted by Ira Glass, produced in collaboration with Chicago Public Media, delivered to stations by PRX The Public Radio Exchange, and has won all of the major broadcasting awards.",
"airtime": "SAT 12pm-1pm, 7pm-8pm",
"imageSrc": "https://ww2.kqed.org/radio/wp-content/uploads/sites/50/2018/04/thisAmericanLife.png",
"officialWebsiteLink": "https://www.thisamericanlife.org/",
"meta": {
"site": "news",
"source": "wbez"
},
"link": "/radio/program/this-american-life",
"subscribe": {
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=201671138&at=11l79Y&ct=nprdirectory",
"rss": "https://www.thisamericanlife.org/podcast/rss.xml"
}
},
"tinydeskradio": {
"id": "tinydeskradio",
"title": "Tiny Desk Radio",
"info": "We're bringing the best of Tiny Desk to the airwaves, only on public radio.",
"airtime": "SUN 8pm and SAT 9pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2025/04/300x300-For-Member-Station-Logo-Tiny-Desk-Radio-@2x.png",
"officialWebsiteLink": "https://www.npr.org/series/g-s1-52030/tiny-desk-radio",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/tinydeskradio",
"subscribe": {
"rss": "https://feeds.npr.org/g-s1-52030/rss.xml"
}
},
"wait-wait-dont-tell-me": {
"id": "wait-wait-dont-tell-me",
"title": "Wait Wait... Don't Tell Me!",
"info": "Peter Sagal and Bill Kurtis host the weekly NPR News quiz show alongside some of the best and brightest news and entertainment personalities.",
"airtime": "SUN 10am-11am, SAT 11am-12pm, SAT 6pm-7pm",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Wait-Wait-Podcast-Tile-300x300-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/wait-wait-dont-tell-me/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/wait-wait-dont-tell-me",
"subscribe": {
"npr": "https://rpb3r.app.goo.gl/Xogv",
"apple": "https://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?s=143441&mt=2&id=121493804&at=11l79Y&ct=nprdirectory",
"tuneIn": "https://tunein.com/radio/Wait-Wait-Dont-Tell-Me-p46/",
"rss": "https://feeds.npr.org/344098539/podcast.xml"
}
},
"weekend-edition-saturday": {
"id": "weekend-edition-saturday",
"title": "Weekend Edition Saturday",
"info": "Weekend Edition Saturday wraps up the week's news and offers a mix of analysis and features on a wide range of topics, including arts, sports, entertainment, and human interest stories. The two-hour program is hosted by NPR's Peabody Award-winning Scott Simon.",
"airtime": "SAT 5am-10am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Weekend-Edition-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/weekend-edition-saturday/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/weekend-edition-saturday"
},
"weekend-edition-sunday": {
"id": "weekend-edition-sunday",
"title": "Weekend Edition Sunday",
"info": "Weekend Edition Sunday features interviews with newsmakers, artists, scientists, politicians, musicians, writers, theologians and historians. The program has covered news events from Nelson Mandela's 1990 release from a South African prison to the capture of Saddam Hussein.",
"airtime": "SUN 5am-10am",
"imageSrc": "https://cdn.kqed.org/wp-content/uploads/2024/04/Weekend-Edition-Podcast-Tile-360x360-1.jpg",
"officialWebsiteLink": "https://www.npr.org/programs/weekend-edition-sunday/",
"meta": {
"site": "news",
"source": "npr"
},
"link": "/radio/program/weekend-edition-sunday"
}
},
"racesReducer": {},
"racesGenElectionReducer": {},
"racesGenElection2026Reducer": {},
"racesGenElection2026PresidentialReducer": {},
"radioSchedulesReducer": {},
"listsReducer": {
"posts/news?tag=chatgpt": {
"isFetching": false,
"latestQuery": {
"from": 0,
"size": 12
},
"vitalsOnly": false,
"totalRequested": 12,
"isLoading": false,
"isLoadingMore": true,
"total": {
"value": 38,
"relation": "eq"
},
"items": [
"news_12096019",
"news_12095554",
"news_12094081",
"news_12092162",
"news_12087559",
"news_12084062",
"news_12083612",
"news_12083278",
"news_12083224",
"news_12082428",
"news_12082064",
"news_12081798"
],
"complete": true
}
},
"recallGuideReducer": {
"intros": {},
"policy": {},
"candidates": {}
},
"savedArticleReducer": {
"articles": [],
"status": {}
},
"newslettersReducer": {
"isFetching": false,
"fetchFailed": false,
"hasFetched": false,
"newsletters": {},
"isSubscribing": false,
"isUnsubscribing": false,
"subscribedNewsletters": {}
},
"termsReducer": {
"about": {
"name": "About",
"type": "terms",
"id": "about",
"slug": "about",
"link": "/about",
"taxonomy": "site"
},
"arts": {
"name": "Arts & Culture",
"grouping": [
"arts",
"pop",
"trulyca"
],
"description": "KQED Arts provides daily in-depth coverage of the Bay Area's music, art, film, performing arts, literature and arts news, as well as cultural commentary and criticism.",
"type": "terms",
"id": "arts",
"slug": "arts",
"link": "/arts",
"taxonomy": "site"
},
"artschool": {
"name": "Art School",
"parent": "arts",
"type": "terms",
"id": "artschool",
"slug": "artschool",
"link": "/artschool",
"taxonomy": "site"
},
"bayareabites": {
"name": "KQED food",
"grouping": [
"food",
"bayareabites",
"checkplease"
],
"parent": "food",
"type": "terms",
"id": "bayareabites",
"slug": "bayareabites",
"link": "/food",
"taxonomy": "site"
},
"bayareahiphop": {
"name": "Bay Area Hiphop",
"type": "terms",
"id": "bayareahiphop",
"slug": "bayareahiphop",
"link": "/bayareahiphop",
"taxonomy": "site"
},
"campaign21": {
"name": "Campaign 21",
"type": "terms",
"id": "campaign21",
"slug": "campaign21",
"link": "/campaign21",
"taxonomy": "site"
},
"careers": {
"name": "Careers",
"type": "terms",
"id": "careers",
"slug": "careers",
"link": "/careers",
"taxonomy": "site"
},
"checkplease": {
"name": "KQED food",
"grouping": [
"food",
"bayareabites",
"checkplease"
],
"parent": "food",
"type": "terms",
"id": "checkplease",
"slug": "checkplease",
"link": "/food",
"taxonomy": "site"
},
"education": {
"name": "Education",
"grouping": [
"education"
],
"type": "terms",
"id": "education",
"slug": "education",
"link": "/education",
"taxonomy": "site"
},
"elections": {
"name": "Elections",
"type": "terms",
"id": "elections",
"slug": "elections",
"link": "/elections",
"taxonomy": "site"
},
"events": {
"name": "Events",
"type": "terms",
"id": "events",
"slug": "events",
"link": "/events",
"taxonomy": "site"
},
"event": {
"name": "Event",
"alias": "events",
"type": "terms",
"id": "event",
"slug": "event",
"link": "/event",
"taxonomy": "site"
},
"filmschoolshorts": {
"name": "Film School Shorts",
"type": "terms",
"id": "filmschoolshorts",
"slug": "filmschoolshorts",
"link": "/filmschoolshorts",
"taxonomy": "site"
},
"food": {
"name": "KQED food",
"grouping": [
"food",
"bayareabites",
"checkplease"
],
"type": "terms",
"id": "food",
"slug": "food",
"link": "/food",
"taxonomy": "site"
},
"forum": {
"name": "Forum",
"relatedContentQuery": "posts/forum?",
"parent": "news",
"type": "terms",
"id": "forum",
"slug": "forum",
"link": "/forum",
"taxonomy": "site"
},
"futureofyou": {
"name": "Future of You",
"grouping": [
"science",
"futureofyou"
],
"parent": "science",
"type": "terms",
"id": "futureofyou",
"slug": "futureofyou",
"link": "/futureofyou",
"taxonomy": "site"
},
"jpepinheart": {
"name": "KQED food",
"relatedContentQuery": "posts/food,bayareabites,checkplease",
"parent": "food",
"type": "terms",
"id": "jpepinheart",
"slug": "jpepinheart",
"link": "/food",
"taxonomy": "site"
},
"liveblog": {
"name": "Live Blog",
"type": "terms",
"id": "liveblog",
"slug": "liveblog",
"link": "/liveblog",
"taxonomy": "site"
},
"livetv": {
"name": "Live TV",
"parent": "tv",
"type": "terms",
"id": "livetv",
"slug": "livetv",
"link": "/livetv",
"taxonomy": "site"
},
"lowdown": {
"name": "The Lowdown",
"relatedContentQuery": "posts/lowdown?",
"parent": "news",
"type": "terms",
"id": "lowdown",
"slug": "lowdown",
"link": "/lowdown",
"taxonomy": "site"
},
"mindshift": {
"name": "Mindshift",
"parent": "news",
"description": "MindShift explores the future of education by highlighting the innovative – and sometimes counterintuitive – ways educators and parents are helping all children succeed.",
"type": "terms",
"id": "mindshift",
"slug": "mindshift",
"link": "/mindshift",
"taxonomy": "site"
},
"news": {
"name": "News",
"grouping": [
"news",
"forum"
],
"type": "terms",
"id": "news",
"slug": "news",
"link": "/news",
"taxonomy": "site"
},
"newsletters": {
"name": "newsletters",
"type": "terms",
"id": "newsletters",
"slug": "newsletters",
"link": "/newsletters",
"taxonomy": "site"
},
"perspectives": {
"name": "Perspectives",
"parent": "radio",
"type": "terms",
"id": "perspectives",
"slug": "perspectives",
"link": "/perspectives",
"taxonomy": "site"
},
"podcasts": {
"name": "Podcasts",
"type": "terms",
"id": "podcasts",
"slug": "podcasts",
"link": "/podcasts",
"taxonomy": "site"
},
"pop": {
"name": "Pop",
"parent": "arts",
"type": "terms",
"id": "pop",
"slug": "pop",
"link": "/pop",
"taxonomy": "site"
},
"pressroom": {
"name": "Pressroom",
"type": "terms",
"id": "pressroom",
"slug": "pressroom",
"link": "/pressroom",
"taxonomy": "site"
},
"quest": {
"name": "Quest",
"parent": "science",
"type": "terms",
"id": "quest",
"slug": "quest",
"link": "/quest",
"taxonomy": "site"
},
"radio": {
"name": "Radio",
"grouping": [
"forum",
"perspectives"
],
"description": "Listen to KQED Public Radio – home of Forum and The California Report – on 88.5 FM in San Francisco, 89.3 FM in Sacramento, 88.3 FM in Santa Rosa and 88.1 FM in Martinez.",
"type": "terms",
"id": "radio",
"slug": "radio",
"link": "/radio",
"taxonomy": "site"
},
"root": {
"name": "KQED",
"image": "https://ww2.kqed.org/app/uploads/2020/02/KQED-OG-Image@1x.png",
"imageWidth": 1200,
"imageHeight": 630,
"headData": {
"title": "KQED | News, Radio, Podcasts, TV | Public Media for Northern California",
"description": "KQED provides public radio, television, and independent reporting on issues that matter to the Bay Area. We’re the NPR and PBS member station for Northern California."
},
"type": "terms",
"id": "root",
"slug": "root",
"link": "/root",
"taxonomy": "site"
},
"science": {
"name": "Science",
"grouping": [
"science",
"futureofyou"
],
"description": "KQED Science brings you award-winning science and environment coverage from the Bay Area and beyond.",
"type": "terms",
"id": "science",
"slug": "science",
"link": "/science",
"taxonomy": "site"
},
"stateofhealth": {
"name": "State of Health",
"parent": "science",
"type": "terms",
"id": "stateofhealth",
"slug": "stateofhealth",
"link": "/stateofhealth",
"taxonomy": "site"
},
"support": {
"name": "Support",
"type": "terms",
"id": "support",
"slug": "support",
"link": "/support",
"taxonomy": "site"
},
"thedolist": {
"name": "The Do List",
"parent": "arts",
"type": "terms",
"id": "thedolist",
"slug": "thedolist",
"link": "/thedolist",
"taxonomy": "site"
},
"trulyca": {
"name": "Truly CA",
"grouping": [
"arts",
"pop",
"trulyca"
],
"parent": "arts",
"type": "terms",
"id": "trulyca",
"slug": "trulyca",
"link": "/trulyca",
"taxonomy": "site"
},
"tv": {
"name": "TV",
"type": "terms",
"id": "tv",
"slug": "tv",
"link": "/tv",
"taxonomy": "site"
},
"voterguide": {
"name": "Voter Guide",
"parent": "elections",
"alias": "elections",
"type": "terms",
"id": "voterguide",
"slug": "voterguide",
"link": "/voterguide",
"taxonomy": "site"
},
"guiaelectoral": {
"name": "Guia Electoral",
"parent": "elections",
"alias": "elections",
"type": "terms",
"id": "guiaelectoral",
"slug": "guiaelectoral",
"link": "/guiaelectoral",
"taxonomy": "site"
},
"news_tag_chatgpt": {
"isLoading": true
},
"news_32668": {
"type": "terms",
"id": "news_32668",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "32668",
"found": true
},
"relationships": {},
"featImg": null,
"name": "ChatGPT",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "ChatGPT Archives | KQED News",
"ogDescription": null
},
"ttid": 32685,
"slug": "chatgpt",
"isLoading": false,
"link": "/news/tag/chatgpt"
},
"source_news_12096019": {
"type": "terms",
"id": "source_news_12096019",
"meta": {
"override": true
},
"name": "The Associated Press",
"isLoading": false
},
"source_news_12094081": {
"type": "terms",
"id": "source_news_12094081",
"meta": {
"override": true
},
"name": "CalMatters",
"link": "https://calmatters.org/",
"isLoading": false
},
"source_news_12087559": {
"type": "terms",
"id": "source_news_12087559",
"meta": {
"override": true
},
"name": "CalMatters",
"link": "https://calmatters.org/",
"isLoading": false
},
"source_news_12082428": {
"type": "terms",
"id": "source_news_12082428",
"meta": {
"override": true
},
"name": "The Bay",
"link": "https://www.kqed.org/podcasts/thebay",
"isLoading": false
},
"news_31795": {
"type": "terms",
"id": "news_31795",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "31795",
"found": true
},
"relationships": {},
"featImg": null,
"name": "California",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "California Archives | KQED News",
"ogDescription": null
},
"ttid": 31812,
"slug": "california",
"isLoading": false,
"link": "/news/category/california"
},
"news_8": {
"type": "terms",
"id": "news_8",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "8",
"found": true
},
"relationships": {},
"featImg": null,
"name": "News",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "News Archives | KQED News",
"ogDescription": null
},
"ttid": 8,
"slug": "news",
"isLoading": false,
"link": "/news/category/news"
},
"news_248": {
"type": "terms",
"id": "news_248",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "248",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Technology",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Technology Archives | KQED News",
"ogDescription": null
},
"ttid": 256,
"slug": "technology",
"isLoading": false,
"link": "/news/category/technology"
},
"news_34755": {
"type": "terms",
"id": "news_34755",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "34755",
"found": true
},
"relationships": {},
"name": "artificial intelligence",
"slug": "artificial-intelligence",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "artificial intelligence | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 34772,
"isLoading": false,
"link": "/news/tag/artificial-intelligence"
},
"news_18538": {
"type": "terms",
"id": "news_18538",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "18538",
"found": true
},
"relationships": {},
"featImg": null,
"name": "California",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "California Archives | KQED News",
"ogDescription": null
},
"ttid": 31,
"slug": "california",
"isLoading": false,
"link": "/news/tag/california"
},
"news_36279": {
"type": "terms",
"id": "news_36279",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "36279",
"found": true
},
"relationships": {},
"name": "chatbot",
"slug": "chatbot",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "chatbot | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 36296,
"isLoading": false,
"link": "/news/tag/chatbot"
},
"news_17996": {
"type": "terms",
"id": "news_17996",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "17996",
"found": true
},
"relationships": {},
"featImg": null,
"name": "News",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "News Archives | KQED News",
"ogDescription": null
},
"ttid": 18030,
"slug": "news",
"isLoading": false,
"link": "/news/tag/news"
},
"news_33542": {
"type": "terms",
"id": "news_33542",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33542",
"found": true
},
"relationships": {},
"featImg": null,
"name": "OpenAI",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "OpenAI Archives | KQED News",
"ogDescription": null
},
"ttid": 33559,
"slug": "openai",
"isLoading": false,
"link": "/news/tag/openai"
},
"news_35836": {
"type": "terms",
"id": "news_35836",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "35836",
"found": true
},
"relationships": {},
"name": "safety",
"slug": "safety",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "safety | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 35853,
"isLoading": false,
"link": "/news/tag/safety"
},
"news_38": {
"type": "terms",
"id": "news_38",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "38",
"found": true
},
"relationships": {},
"featImg": null,
"name": "San Francisco",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "San Francisco Archives | KQED News",
"ogDescription": null
},
"ttid": 58,
"slug": "san-francisco",
"isLoading": false,
"link": "/news/tag/san-francisco"
},
"news_35940": {
"type": "terms",
"id": "news_35940",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "35940",
"found": true
},
"relationships": {},
"name": "tech industry",
"slug": "tech-industry",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "tech industry | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 35957,
"isLoading": false,
"link": "/news/tag/tech-industry"
},
"news_33729": {
"type": "terms",
"id": "news_33729",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33729",
"found": true
},
"relationships": {},
"featImg": null,
"name": "San Francisco",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "San Francisco Archives | KQED News",
"ogDescription": null
},
"ttid": 33746,
"slug": "san-francisco",
"isLoading": false,
"link": "/news/interest/san-francisco"
},
"news_33737": {
"type": "terms",
"id": "news_33737",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33737",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Science",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Science Archives | KQED News",
"ogDescription": null
},
"ttid": 33754,
"slug": "science",
"isLoading": false,
"link": "/news/interest/science"
},
"news_33732": {
"type": "terms",
"id": "news_33732",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33732",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Technology",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Technology Archives | KQED News",
"ogDescription": null
},
"ttid": 33749,
"slug": "technology",
"isLoading": false,
"link": "/news/interest/technology"
},
"news_18540": {
"type": "terms",
"id": "news_18540",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "18540",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Education",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Education Archives | KQED News",
"ogDescription": null
},
"ttid": 2595,
"slug": "education",
"isLoading": false,
"link": "/news/category/education"
},
"news_25184": {
"type": "terms",
"id": "news_25184",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "25184",
"found": true
},
"relationships": {},
"featImg": null,
"name": "AI",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "AI Archives | KQED News",
"ogDescription": null
},
"ttid": 25201,
"slug": "ai",
"isLoading": false,
"link": "/news/tag/ai"
},
"news_20013": {
"type": "terms",
"id": "news_20013",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "20013",
"found": true
},
"relationships": {},
"featImg": null,
"name": "education",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "education Archives | KQED News",
"ogDescription": null
},
"ttid": 20030,
"slug": "education",
"isLoading": false,
"link": "/news/tag/education"
},
"news_27626": {
"type": "terms",
"id": "news_27626",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "27626",
"found": true
},
"relationships": {},
"featImg": null,
"name": "featured-news",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "featured-news Archives | KQED News",
"ogDescription": null
},
"ttid": 27643,
"slug": "featured-news",
"isLoading": false,
"link": "/news/tag/featured-news"
},
"news_22602": {
"type": "terms",
"id": "news_22602",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "22602",
"found": true
},
"relationships": {},
"featImg": null,
"name": "school",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "school Archives | KQED News",
"ogDescription": null
},
"ttid": 22619,
"slug": "school",
"isLoading": false,
"link": "/news/tag/school"
},
"news_3457": {
"type": "terms",
"id": "news_3457",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "3457",
"found": true
},
"relationships": {},
"name": "Students",
"slug": "students",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "Students | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 3475,
"isLoading": false,
"link": "/news/tag/students"
},
"news_348": {
"type": "terms",
"id": "news_348",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "348",
"found": true
},
"relationships": {},
"featImg": null,
"name": "surveys",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "surveys Archives | KQED News",
"ogDescription": null
},
"ttid": 356,
"slug": "surveys",
"isLoading": false,
"link": "/news/tag/surveys"
},
"news_1631": {
"type": "terms",
"id": "news_1631",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "1631",
"found": true
},
"relationships": {},
"name": "Technology",
"slug": "technology",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "Technology | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 1643,
"isLoading": false,
"link": "/news/tag/technology"
},
"news_20385": {
"type": "terms",
"id": "news_20385",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "20385",
"found": true
},
"relationships": {},
"featImg": null,
"name": "teens",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "teens Archives | KQED News",
"ogDescription": null
},
"ttid": 20402,
"slug": "teens",
"isLoading": false,
"link": "/news/tag/teens"
},
"news_98": {
"type": "terms",
"id": "news_98",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "98",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Youth",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Youth Archives | KQED News",
"ogDescription": null
},
"ttid": 101,
"slug": "youth",
"isLoading": false,
"link": "/news/tag/youth"
},
"news_33746": {
"type": "terms",
"id": "news_33746",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33746",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Education",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Education Archives | KQED News",
"ogDescription": null
},
"ttid": 33763,
"slug": "education",
"isLoading": false,
"link": "/news/interest/education"
},
"news_33733": {
"type": "terms",
"id": "news_33733",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33733",
"found": true
},
"relationships": {},
"featImg": null,
"name": "News",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "News Archives | KQED News",
"ogDescription": null
},
"ttid": 33750,
"slug": "news",
"isLoading": false,
"link": "/news/interest/news"
},
"news_457": {
"type": "terms",
"id": "news_457",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "457",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Health",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Health Archives | KQED News",
"ogDescription": null
},
"ttid": 16998,
"slug": "health",
"isLoading": false,
"link": "/news/category/health"
},
"news_421": {
"type": "terms",
"id": "news_421",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "421",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Kaiser Permanente",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Kaiser Permanente Archives | KQED News",
"ogDescription": null
},
"ttid": 430,
"slug": "kaiser-permanente",
"isLoading": false,
"link": "/news/tag/kaiser-permanente"
},
"news_2109": {
"type": "terms",
"id": "news_2109",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "2109",
"found": true
},
"relationships": {},
"featImg": null,
"name": "mental health",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "mental health Archives | KQED News",
"ogDescription": null
},
"ttid": 2124,
"slug": "mental-health",
"isLoading": false,
"link": "/news/tag/mental-health"
},
"news_20782": {
"type": "terms",
"id": "news_20782",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "20782",
"found": true
},
"relationships": {},
"featImg": null,
"name": "therapy",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "therapy Archives | KQED News",
"ogDescription": null
},
"ttid": 20799,
"slug": "therapy",
"isLoading": false,
"link": "/news/tag/therapy"
},
"news_18481": {
"type": "terms",
"id": "news_18481",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "18481",
"found": true
},
"relationships": {},
"featImg": null,
"name": "CALmatters",
"description": null,
"taxonomy": "affiliate",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "CALmatters Archives | KQED Arts",
"ogDescription": null
},
"ttid": 18515,
"slug": "calmatters",
"isLoading": false,
"link": "/news/affiliate/calmatters"
},
"news_33747": {
"type": "terms",
"id": "news_33747",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33747",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Health",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Health Archives | KQED News",
"ogDescription": null
},
"ttid": 33764,
"slug": "health",
"isLoading": false,
"link": "/news/interest/health"
},
"news_28250": {
"type": "terms",
"id": "news_28250",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "28250",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Local",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Local Archives | KQED News",
"ogDescription": null
},
"ttid": 28267,
"slug": "local",
"isLoading": false,
"link": "/news/category/local"
},
"news_35758": {
"type": "terms",
"id": "news_35758",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "35758",
"found": true
},
"relationships": {},
"name": "Open AI",
"slug": "open-ai",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "Open AI | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 35775,
"isLoading": false,
"link": "/news/tag/open-ai"
},
"news_34586": {
"type": "terms",
"id": "news_34586",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "34586",
"found": true
},
"relationships": {},
"name": "Silicon Valley",
"slug": "silicon-valley",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "Silicon Valley | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 34603,
"isLoading": false,
"link": "/news/tag/silicon-valley"
},
"news_16": {
"type": "terms",
"id": "news_16",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "16",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Gavin Newsom",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Gavin Newsom Archives | KQED News",
"ogDescription": null
},
"ttid": 16,
"slug": "gavin-newsom",
"isLoading": false,
"link": "/news/tag/gavin-newsom"
},
"news_33738": {
"type": "terms",
"id": "news_33738",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33738",
"found": true
},
"relationships": {},
"featImg": null,
"name": "California",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "California Archives | KQED News",
"ogDescription": null
},
"ttid": 33755,
"slug": "california",
"isLoading": false,
"link": "/news/interest/california"
},
"news_6188": {
"type": "terms",
"id": "news_6188",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "6188",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Law and Justice",
"description": null,
"taxonomy": "category",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Law and Justice Archives | KQED News",
"ogDescription": null
},
"ttid": 6212,
"slug": "law-and-justice",
"isLoading": false,
"link": "/news/category/law-and-justice"
},
"news_1386": {
"type": "terms",
"id": "news_1386",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "1386",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Bay Area",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Bay Area Archives | KQED News",
"ogDescription": null
},
"ttid": 1398,
"slug": "bay-area",
"isLoading": false,
"link": "/news/tag/bay-area"
},
"news_3897": {
"type": "terms",
"id": "news_3897",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "3897",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Elon Musk",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Elon Musk Archives | KQED News",
"ogDescription": null
},
"ttid": 3916,
"slug": "elon-musk",
"isLoading": false,
"link": "/news/tag/elon-musk"
},
"news_34054": {
"type": "terms",
"id": "news_34054",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "34054",
"found": true
},
"relationships": {},
"featImg": null,
"name": "oakland",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "oakland Archives | KQED News",
"ogDescription": null
},
"ttid": 34071,
"slug": "oakland",
"isLoading": false,
"link": "/news/tag/oakland"
},
"news_33543": {
"type": "terms",
"id": "news_33543",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33543",
"found": true
},
"relationships": {},
"name": "Sam Altman",
"slug": "sam-altman",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "Sam Altman | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null,
"metaRobotsNoIndex": "noindex"
},
"ttid": 33560,
"isLoading": false,
"link": "/news/tag/sam-altman"
},
"news_33730": {
"type": "terms",
"id": "news_33730",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33730",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Oakland",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Oakland Archives | KQED News",
"ogDescription": null
},
"ttid": 33747,
"slug": "oakland",
"isLoading": false,
"link": "/news/interest/oakland"
},
"news_21891": {
"type": "terms",
"id": "news_21891",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "21891",
"found": true
},
"relationships": {},
"featImg": null,
"name": "lawsuits",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "lawsuits Archives | KQED News",
"ogDescription": null
},
"ttid": 21908,
"slug": "lawsuits",
"isLoading": false,
"link": "/news/tag/lawsuits"
},
"news_19954": {
"type": "terms",
"id": "news_19954",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "19954",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Law and Justice",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Law and Justice Archives | KQED News",
"ogDescription": null
},
"ttid": 19971,
"slug": "law-and-justice",
"isLoading": false,
"link": "/news/tag/law-and-justice"
},
"news_36810": {
"type": "terms",
"id": "news_36810",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "36810",
"found": true
},
"relationships": {},
"name": "federal trial",
"slug": "federal-trial",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "federal trial | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 36827,
"isLoading": false,
"link": "/news/tag/federal-trial"
},
"news_33812": {
"type": "terms",
"id": "news_33812",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33812",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Interests",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Interests Archives | KQED News",
"ogDescription": null
},
"ttid": 33829,
"slug": "interests",
"isLoading": false,
"link": "/news/tag/interests"
},
"news_22598": {
"type": "terms",
"id": "news_22598",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "22598",
"found": true
},
"relationships": {},
"featImg": null,
"name": "The Bay",
"description": "\u003cimg class=\"alignnone size-medium wp-image-11638190\" src=\"https://ww2.kqed.org/news/wp-content/uploads/sites/10/2018/02/TheBay_1200x6301.png\" alt=\"\" />\r\n\u003cbr/>\r\n\r\nEvery good story starts local. So that’s where we start. \u003ci>The Bay\u003c/i> is storytelling for daily news. KQED host Devin Katayama talks with reporters to help us make sense of what’s happening in the Bay Area. One story. One conversation. One idea.\r\n\r\n\u003cstrong>Subscribe to The Bay:\u003c/strong>\r\n\r\n\u003ca href=\"https://itunes.apple.com/us/podcast/the-bay/id1350043452?mt=2\">\u003cimg src=\"https://ww2.kqed.org/news/wp-content/uploads/sites/10/2018/01/Listen_on_Apple_Podcasts_sRGB_US-e1515635079510.png\" />\u003c/a>",
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": "Every good story starts local. So that’s where we start. The Bay is storytelling for daily news. KQED host Devin Katayama talks with reporters to help us make sense of what’s happening in the Bay Area. One story. One conversation. One idea. Subscribe to The Bay:",
"title": "The Bay Archives | KQED News",
"ogDescription": null
},
"ttid": 22615,
"slug": "the-bay",
"isLoading": false,
"link": "/news/tag/the-bay"
},
"news_34167": {
"type": "terms",
"id": "news_34167",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "34167",
"found": true
},
"relationships": {},
"name": "Criminal Justice",
"slug": "criminal-justice",
"taxonomy": "category",
"description": null,
"featImg": null,
"headData": {
"title": "Criminal Justice Archives | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 34184,
"isLoading": false,
"link": "/news/category/criminal-justice"
},
"news_17725": {
"type": "terms",
"id": "news_17725",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "17725",
"found": true
},
"relationships": {},
"featImg": null,
"name": "criminal justice",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "criminal justice Archives | KQED News",
"ogDescription": null
},
"ttid": 17759,
"slug": "criminal-justice",
"isLoading": false,
"link": "/news/tag/criminal-justice"
},
"news_22434": {
"type": "terms",
"id": "news_22434",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "22434",
"found": true
},
"relationships": {},
"featImg": null,
"name": "death",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "death Archives | KQED News",
"ogDescription": null
},
"ttid": 22451,
"slug": "death",
"isLoading": false,
"link": "/news/tag/death"
},
"news_35784": {
"type": "terms",
"id": "news_35784",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "35784",
"found": true
},
"relationships": {},
"name": "gun violence",
"slug": "gun-violence",
"taxonomy": "tag",
"description": null,
"featImg": null,
"headData": {
"title": "gun violence | KQED News",
"description": null,
"ogTitle": null,
"ogDescription": null,
"ogImgId": null,
"twTitle": null,
"twDescription": null,
"twImgId": null
},
"ttid": 35801,
"isLoading": false,
"link": "/news/tag/gun-violence"
},
"news_33745": {
"type": "terms",
"id": "news_33745",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "33745",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Criminal Justice",
"description": null,
"taxonomy": "interest",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Criminal Justice Archives | KQED News",
"ogDescription": null
},
"ttid": 33762,
"slug": "criminal-justice",
"isLoading": false,
"link": "/news/interest/criminal-justice"
},
"news_57": {
"type": "terms",
"id": "news_57",
"meta": {
"index": "terms_1716263798",
"site": "news",
"id": "57",
"found": true
},
"relationships": {},
"featImg": null,
"name": "Tesla",
"description": null,
"taxonomy": "tag",
"headData": {
"twImgId": null,
"twTitle": null,
"ogTitle": null,
"ogImgId": null,
"twDescription": null,
"description": null,
"title": "Tesla Archives | KQED News",
"ogDescription": null
},
"ttid": 57,
"slug": "tesla",
"isLoading": false,
"link": "/news/tag/tesla"
}
},
"userPermissionsReducer": {
"wpLoggedIn": false
},
"eventsReducer": {},
"fssReducer": {},
"tvDailyScheduleReducer": {},
"tvWeeklyScheduleReducer": {},
"tvPrimetimeScheduleReducer": {},
"tvMonthlyScheduleReducer": {},
"userAccountReducer": {
"user": {
"email": null,
"emailStatus": "EMAIL_UNVALIDATED",
"loggedStatus": "LOGGED_OUT",
"loggingChecked": false,
"articles": [],
"firstName": null,
"lastName": null,
"phoneNumber": null,
"fetchingMembership": false,
"membershipError": false,
"memberships": [
{
"id": null,
"startDate": null,
"firstName": null,
"lastName": null,
"familyNumber": null,
"memberNumber": null,
"memberSince": null,
"expirationDate": null,
"pfsEligible": false,
"isSustaining": false,
"membershipLevel": "Prospect",
"membershipStatus": "Non Member",
"lastGiftDate": null,
"renewalDate": null,
"lastDonationAmount": null
}
]
},
"authModal": {
"isOpen": false,
"view": "LANDING_VIEW"
},
"error": null
},
"youthMediaReducer": {},
"checkPleaseReducer": {
"filterData": {
"region": {
"key": "Restaurant Region",
"filters": [
"Any Region"
]
},
"cuisine": {
"key": "Restaurant Cuisine",
"filters": [
"Any Cuisine"
]
}
},
"restaurantDataById": {},
"restaurantIdsSorted": [],
"error": null
},
"userAgentReducer": {
"userAgent": "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; ClaudeBot/1.0; +claudebot@anthropic.com)",
"isBot": true
},
"requestOutcomesReducer": {
"notFound": []
}
}
