‘The AI Doc’ Is Probably the Scariest Movie You’ll See All Year

I’ve been asking my pregnant friends the same aghast question for about 20 years now: “Aren’t you worried about climate change?”

More recently, that query has morphed into a far more frenzied: “Aren’t you worried about climate change, and fascism, and the robots that are already plotting to kill us?”

As soon as his wife got pregnant a few years ago, Daniel Roher started worrying about those things as well — particularly that last one. Rather than simply collapsing into despair, Roher channeled his anxiety into thoroughly exploring the myths and realities of AI as we currently know it. The resulting documentary, The AI Doc: Or How I Became an Apocaloptimist, consults a plethora of experts on the topic, with often confusing results.

Roher gathers the interviewees into three broad groups: the terrifyingly pessimistic ones, the naively optimistic ones and the CEOs who are casually working on something that may or may not spark humanity’s demise. (Well, three out of five of them, anyway: OpenAI CEO Sam Altman, Google DeepMind CEO Demis Hassabis and Anthropic CEO Dario Amodei all appear. Mark Zuckerberg declined to participate and Elon Musk apparently backed out at the last minute.)

The worst AI predictions are presented first. Eliezer Yudkowsky, co-founder of the Machine Intelligence Research Institute, calmly talks of the “abrupt extermination” of humanity. Author and historian Yuval Noah Harari calls AI “a deadly threat.”

Center for Humane Technology President Tristan Harris — one of the most measured commentators in the movie — also shares some truly sobering views, the worst of which is that he knows active AI researchers who “don’t expect their children to make it to high school.” It doesn’t help matters that machine learning researcher Shane Legg follows this with the assertion, “The really powerful systems are coming and they’re coming soon — this is all just a warm-up.”

When the AI optimists arrive, it’s to assert that smart machines might one day save humanity from disease, climate change and asteroids. Canadian physicist Guillaume Verdon is the perkiest of them all, calling our present era “a glorious period of human transformation.” While the pessimists suggest that super-intelligent machines are going to end civilization as we know it, the optimists are convinced that humanity won’t survive without the assistance of AI that is smarter than us. (Roher rightfully wonders aloud at this stage how AI is going to save us from climate change when data centers pose such well-established threats to ecosystems.)

When Roher asks Aza Raskin, co-founder of the Center for Humane Technology and Earth Species Project, who’s closer to the truth — the pessimists or the optimists — Raskin offers the frustratingly opaque: “They’re both right and neither side goes far enough.”

The AI Doc, then, is an appropriately confounding documentary for a labyrinthian topic that is, at its core, too immense for most laypeople to fully grapple with yet. Roher makes for a relatable everyman throughout this rollercoaster of dread, relief, hope and fear.

By the time the AI CEOs arrive — they show up late in the picture and are, mercifully, not permitted to dominate the conversation — the film mostly confirms that predictions for AI’s endgame lie entirely in the eye of the beholder. It’s certainly interesting that each CEO seems to believe his company is more morally sound than the next.

In his conversations with Roher, Sam Altman talks a good game about the safety protocols that OpenAI has in place. Given his company’s highly controversial new contract with the Department of Defense, his words will either ring hollow or serve as comfort, depending on your viewpoint. For his part in The AI Doc, Anthropic’s Dario Amodei simply says, “Am I confident that everything’s going to work out? No, I’m not.” Hassabis is even more vague: “If something is possible to do, humanity is going to do it,” he says.

In the end, Roher’s foremost question in the documentary — “Is now a terrible time to have a kid?” — does get an answer of sorts. That, broadly, is a resounding “Yes, but…” The “but” here is that it’s probably never been a great time to have kids, so why stop now? Roher’s own parents try to comfort him with the fact that they started their own family in the midst of the Cold War, when the threat of nuclear annihilation felt ever-present.

If Roher started the project seeking comfort, you can rest (un)assured after viewing The AI Doc that Roher did not find what he was looking for — not in any real, permanent sense, anyway. It’s unlikely that most viewers will reach any solid conclusions either. The feeling I left The AI Doc with is that the future of AI is overwhelmingly — and unfortunately — out of the hands of everyday people. That we’re being forced to put our faith in tech executives to make the right decisions is probably the scariest thing of all.

‘The AI Doc: Or How I Became An Apocaloptimist’ is released nationwide on March 27, 2026.