Adults Have Struggled to Set Rules for AI In School. These Teens Figured It Out

BOSTON — Over the course of one frantic, late-July weekend, 98 teens representing all 50 states put their busy lives on hold to try to do something remarkable.

The students, most of them high school juniors and seniors, gathered in a precise replica of the U.S. Senate chamber and attempted what the real, adult Congress has largely failed to do: draft and pass “legislation” to help the nation’s K-12 schools navigate the questions and chaos created by artificial intelligence.

Should students be allowed to use AI on assignments? What about on tests? Is it schools’ job to teach the nation’s children what AI is and how to use it? There is no national policy for AI use in schools, meaning states and school districts have had to scramble to craft their own.

“This is a once-in-a-lifetime opportunity for us to actually help and try to change the country for the better,” said student Gabriel Sutphin, representing South Carolina, speaking from the chamber floor at the Edward M. Kennedy Institute for the United States Senate.

To these teens, the stakes for getting ahead of AI are clear. They pointed out how adult leaders were slow to protect students from the risks of other tech developments.

“We missed the mark with social media and phones,” said Hunter Wurzel, a high school senior representing Ohio. “That’s been a huge talking point already [this weekend], is how we failed to implement those things into education and teach the youth how to use them.“

Wurzel and his 97 colleagues were handpicked by their school leaders and hosted by a handful of education- and tech-focused groups, including the Edward M. Kennedy Institute, AASA (The School Superintendents Association), and Day of AI, a nonprofit launched out of MIT’s RAISE AI literacy initiative.

At times during the weekend, where NPR was able to watch the events unfold, the bill seemed doomed, only to be revived by a flurry of student-led negotiation, old-fashioned compromises and some spirited speeches.

“[This bill] is about who gets a seat at the table. Today, students have that seat,” Bena Mejia Romero of Connecticut told her fellow senators before the final vote.

Ultimately, their Students First Act passed overwhelmingly, 82-16.

Among the bill’s requirements: States should provide AI literacy as soon as students get access to classroom devices; AI use should be banned on all graded tests; and teachers can require an oral defense of any project from a student suspected of using AI inappropriately. (You can read the full bill at the end of this article.)

“The adults themselves haven’t been able to come to consensus on this, right? And so now it falls to the kids. Let them lead the way,” said Jeffrey Riley, the executive director of Day of AI and former education commissioner of Massachusetts.

While the Students First Act isn’t actually law, any more than the teens are actually senators, AASA plans to send the text to its members: 10,000 school leaders across the country, many of whom are right now crafting their own AI rules in the absence of regulation.

Rules for students

When students arrived at the start of the weekend, they were presented with a stew of policies that they had suggested in the run-up to the event. Then they were divided into subcommittees — focused on students, educators, administrators and families — to debate the dozens of proposals and create a coherent bill they could all support.

The student-focused section of the final bill includes 15 provisions, several focused on a concern many of the senators shared: that teens are using AI to do their thinking for them.

Without guidance and guardrails, “[AI use] is just gonna go unhinged,” said Campbell McHenry, a high school junior from Mississippi. “Everyone’s gonna stop learning and start relying on AI. Instead of using it as a tool, they’re gonna use it as a brain.”

Among the bill’s student-focused proposals:

• Approved uses: Students should not be allowed to use AI for writing, “though editing, brainstorming and studying” with AI should be allowed after eighth grade.

• Prove mastery: Students permitted to use AI should have to cite how they used it and demonstrate their grasp of a subject through discussions, handwritten tests or oral defenses.

• Plagiarism review: Any student suspected of using AI inappropriately should have their work reviewed not just by AI checker software, which can be unreliable, but also two school officials.

One of the most contentious issues was deciding when students should first be exposed to AI in the classroom: High school? Middle school? Earlier?

Lucia Herrera, a Missouri high schooler, took issue with an early draft of the bill.

“There’s nothing that says students need to be proficient before using it,” Herrera told her subcommittee colleagues. “It’s basically saying they can use it whenever. Which I think is a misuse of it.“

Her subcommittee’s proposed solution: require schools to teach AI literacy as soon as devices are introduced in the classroom, including misinformation, plagiarism, bias in AI design, student privacy, appropriate academic use and AI’s environmental impact.

When the provision came up later that day, before the full Senate, Connecticut’s Bena Mejia Romero told her fellow student lawmakers they should support the new requirement. “If we’re going to be using devices, we should know the dangers and the benefits that they have.”

But one of the senators from California, Om Lala, argued that requiring schools to begin teaching AI literacy would be a heavy lift. “Teachers already are pretty under-resourced.”

Tristan Tjetjep of Tennessee added that he worried how “underprivileged schools” would pay for this new instruction.

Suddenly, the students had run headlong into U.S. schools’ funding inequities — a problem they attempted to resolve with a requirement that the federal government chip in funding.

Rules for educators, parents and caregivers

The Students First Act not only covers students’ use of AI but also insists on rules for educators:

• Approved teacher use of AI: Allowed for tasks including lesson planning, making practice materials, providing feedback and improving instruction.

• Must have a clear AI policy: Each semester, teachers need to post “a clear, written AI usage policy,” explaining when AI is allowed and when it’s not.

What’s more, if an AI detector flags potential inappropriate AI use in a student paper, the bill says “teachers must personally investigate the issue before reporting the student to admin.”

The student senators struggled with the 10 provisions in the bill’s parents/guardians section because, as Isabella Stringer of New Hampshire said on the Senate floor, “We cannot legislate parenthood. Schools cannot tell parents what or how to teach your children.“

As such, some of the provisions here suggest rather than tell. For instance, one says simply that parents should encourage balanced technology habits and emphasize the importance of ethical academic behavior at home.

The early version of the bill also gave parents a full veto over their students’ AI access, but many of the senators argued that AI is here to stay and students must be taught how to use it.

“Giving [parents] the complete say, especially in middle to high school, in what their child can and can’t do with AI is gonna limit [students] because a lot of [adults] don’t even know what AI is themselves,” said junior Tatiyana Reaves of North Carolina.

In one of the weekend’s more passionate exchanges, senators debated what should happen if a student attempts to use a school-based AI tool, like a chatbot, “for emotional support.”

Several teens worried that if a student confided sensitive information about, say, their sexuality or a mental health crisis, they should be able to trust that it would be kept private.

But Dylan Dornack of Iowa argued on the Senate floor that, for students in crisis or experiencing suicidal thoughts, the technology must alert a human.

“This could be lifesaving,” Dornack told his colleagues, “whether or not that student is fearful that they might get outed. This is just one of those things that has to be taken to trusted adults.”

Ultimately, the students agreed: When a school chatbot detects a student in need of emotional support, it should not fulfill the student’s request but instead alert school officials.

The rush to passage

At the end of the whirlwind weekend, the student senators gathered in the replica chamber and more than half volunteered to address their colleagues for one minute each.

Dornack was the first to speak and called the bill an important, if imperfect, foundation for schools’ AI policies. “You have to lay the blocks before you can have the building. So here I say, let’s lay the blocks and let’s pass this.”

But Jadelynn Petitjean of Alabama called the bill an “unstable foundation” and opposed it.

Earlier in the day, West Virginia’s Ethan Liu had also opposed it, highlighting several contradictions in the bill, including amendments that both restricted students’ AI access while appearing to give teachers wide discretion over when and how AI is used.

“I thought it reflected really badly on us. I had a whole different speech planned out against this bill, actually. It was really good,” Liu told the Senate chamber before the final vote.

But at the last minute, the event’s organizers allowed the students to iron out a few of the bill’s inconsistencies, and, as a result, Liu changed his vote, saying “this bill is in much better shape.”

Student after student walked to the microphone at the front of the chamber, and most endorsed the bill, including Kendall Terao-Toma of Hawaii:

“This is for our future — not just for education but for the workforce we will join, the economy that we will build and the communities we will lead.”

Edited by: Nirvi Shah and Nicole Cohen; audio story produced by: Lauren Migaki
Visual design and development by: LA Johnson

Read the full text of the Students First Act: