In “The AI Doc: Or How I Became an Apocaloptimist” (in theaters), Daniel Roher, who directs along with Charlie Tyrell, tells us from the start that he’s as confused as we are about artificial intelligence. But like many people, it was the impending birth of his child that brought him nose to nose with life’s biggest questions, including the one that made him want to make this film: Was he crazy to bring a human into a world that might soon be overrun by machines, or overheated by data centers, or impoverished by wealth gaps brought on by an A.I. labor revolution? Is any of this even going to happen, or is artificial intelligence going to make everything better?
In other words, as the movie’s opening narration puts it: Was this the apocalypse, or did he actually have a reason to be optimistic?
Though the word “apocalypse” is often applied to destruction and chaos, its meaning is rooted in the idea of a moment of unveiling or revelation. So in asking the question, Roher unwittingly hits on something important: Whether it’s nightmarish or utopian, A.I. reveals what we believe about the elements that make a good society, the nature of intelligence, the purpose of work and creativity and the very essence of being human.
And those questions are, indeed, what Roher (who won an Oscar for his documentary “Navalny”) sets out to explore in “The AI Doc,” with himself as wondering inquisitor. He sets up a studio and brings in all kinds of experts, ranging from enthusiastic boosters to deep skeptics and pessimists. He asks them to explain A.I. and, specifically, A.G.I. — artificial general intelligence, the theoretical form of A.I. that would perform all the tasks people can, but better. There are techno-optimists (Jensen Huang of Nvidia, Peter Diamandis of the XPrize Foundation) and techno-doomers (like Eliezer Yudkowsky, who terms himself “the original doom guy”). Dozens of names you recognize if you’ve spent any time reading the tech press — and other names you don’t — discuss their ideas about the future.
The result is less clarifying than bewildering, though it’s often very interesting. The techno-boosters, for instance, don’t often talk about who owns the A.I. technology that might provide health care or tutors; it’s a little terrifying to imagine expanding access to memory in our brains by connecting them to the cloud, as someone suggests in passing. A group of scholars and journalists brings context and clarity, but the film can only do so much in its running time.
The most fascinating part of “The AI Doc” comes when Roher manages to get three A.I. company chief executives to show up for an interview: OpenAI’s Sam Altman, Anthropic’s Dario Amodei and Google DeepMind’s Demis Hassabis. (Mark Zuckerberg and Elon Musk declined to participate.) Each comes across as thoughtful and candid — some more candid than others — and gives you a sense of, for instance, what they think about rushing safety testing measures or working with the government, matters that have become only more important in recent months. Because the question of who owns the technology is nearly as important as the technology itself, I found myself wishing that there were much more of this, and that they were interrogated more carefully, their claims held up to the light with a little more time for contemplation.
But the format of “The AI Doc” was bound from the start to be unsatisfying. The cultural conversation around A.I. became a mess almost as soon as it started: Some people equate A.I. with chatbots like ChatGPT and refuse to learn anything more; others consider it to be the inevitable wave of the future and unquestioningly bend the knee; still others develop and use it no matter the cost to the environment or to human societies. Chief executives have plenty of reasons to make broad claims, and pessimists have ample cause to make bleak declarations. The idea of “apocaloptomism” doesn’t really address any of them. There’s some value, though, in sorting through what’s being said. Collecting bites of it in one place to muse upon is at least somewhere to start.
