A new documentary about artificial intelligence puts CEOs in a tough spot, but it goes a long way with them

It’s not easy interviewing Sam Altman, just ask Adam Bhalla Love, the director of the recent documentary Deepfakes Sam Altman.

Lough originally planned a feature exploring the potential and risks of artificial intelligence that would focus on a conversation with OpenAI’s CEO. But after his inquiries were ignored for months, he instead chose to create a chatbot that mimicked Altman’s speech patterns and approximated his facial expressions via a digital avatar.

However, the real Altman sat down for the new feature Artificial Intelligence Doc: Or How I Became an Apocalypsewhich will be in theaters on March 27. So did Dario Amodei, CEO of Anthropic, and Demis Hassabis, co-founder and CEO of Google’s DeepMind Technologies. (Although the filmmakers said they requested interviews with Meta’s Mark Zuckerberg and X’s Elon Musk, neither showed up.)

It’s an impressive level of access for co-director and star of the documentary Daniel Rohr, whose 2022 documentary NavalnyA film about Russian opposition leader Alexei Navalny, which won an Oscar. The problem is that once they are on camera, Altman et al. Say little we haven’t heard before – and they skate by with off-the-cuff answers regarding their responsibilities to the rest of their species. When Rohr asks Altman why anyone should trust him to direct the rapid acceleration of artificial intelligence, given its extreme consequences, Altman replies: “You shouldn’t.” The line of questioning ends there.

Amnesty International document The film revolves around Rohr’s anxiety over the imminent arrival of his son and first child with his wife, director Caroline Linde. He wonders what kind of world his son will inherit and whether the rise of artificial intelligence will prevent the experiments that develop us into self-sufficient adults. In Rohr’s first several interviews, his worst fears seem to have been confirmed. Tristan Harris, co-founder of the nonprofit Center for Humane Technology, delivers one of the worst punches: “I know people who work in AI who take risks and don’t expect their kids to make it to high school,” conjuring up a scenario in which technology destroys the infrastructure of traditional education.

Despite the growing sense of panic, Rohr and his collaborator Charlie Tyrrell deliver an impressive crash course in artificial intelligence and the biggest questions it poses, aided by Rohr’s insistence on defining terms in plain language rather than emerging buzzwords. Visually, the film is charmingly human, featuring Rohr’s color drawings and paintings, while bizarre stop-motion scenes nod to the influence of producer Daniel Kwan, the Oscar-winning co-director of Everything everywhere at once. Vibrant creativity amid the harbingers of doom provides some hope that Rohr desperately seeks.

However, subsequent interviews with Silicon Valley tech optimists who promised AI that would conquer disease and climate change — followed by CEOs striking their usual balance between hype and tones of sober caution — passed without much questioning of the grandiose claims. We hardly spend a moment thinking about why or how we should expect the current crop of huge, fallible linguistic models to give rise to a mythical “artificial general intelligence” that may surpass human cognition. There are, at best, euphemisms (from venture capitalist Reed Hoffman, for example) that any benefits will come with unspecified harms.

Even when the big players say that the near-term implications of AI are as important as the advent of nuclear weapons, they stick to familiar rules of the game, presenting their products as uniquely important in one way or another — only hinting at that. they They can be trusted to advance them.