The AI Doc: Or How I Became an Apocaloptimist
Directed by Daniel Roher and Charlie Tyrell
Featuring Daniel Roher, Caroline Lindy and Sam Altman
Classification PG; 104 minutes
Opens in theatres March 27
Open this photo in gallery:
A still from The AI Doc: Or How I Became an Apocaloptimist. The documentary includes an interview with OpenAI’s Sam Altman.Courtesy of Focus Features./Supplied
For less than half a second, I thought I’d be semi-clever and use ChatGPT to compose this review of a new documentary about the rise of artificial intelligence. Or maybe I’d go the multimedia route and use Nano Banana to produce an animated critique featuring the likenesses of various copyrighted characters and/or a digitally manipulated visage of the long-dead Roger Ebert. But then my entire being, my soul, violently reacted to such a notion, and I simply put digital pen to paper to reckon with The AI Doc : Or How I Became an Apocaloptimist. Surely the film’s co-director Daniel Roher, the Canadian behind the Oscar-winning doc Navalny, would approve. Or would he?
While the new doc was spurred by Roher’s own existential anxiety about what kind of AI-dominated world he would be bringing his unborn son into, the resulting film feels so determined to walk the middle road between doom times and boom times (hence its cheeky title) that its message cannot help but land as something almost algorithmically mushy.
Roher and his co-director, Charlie Tyrell, are committed to approaching the vastly complicated subject of AI with an open and even-handed perspective. But too many times, this well-intentioned impartiality allows the film’s interview subjects – some of the most powerful, or soon-to-be-powerful, people on the planet – to act as mere unchecked pitchmen. And when the product they are selling amounts to the end of the world as we know it, the time has come to abandon any pretense of polite objectivity and ask the hard, tough, why-on-earth-are-you-doing-this questions.
Why one researcher still views AI as an existential threat to humanity
Then again, maybe I’m falling right into the film’s trap, which from the start announces its intent to step away from the extreme reactions AI has engendered in order to find out just what the heck it is that has gotten everyone so upset. To that end, Roher and Tyrell craft an enjoyably breezy, frequently witty “AI 101” primer.
Narrated first-person style by Roher, whose warm and inviting tone is occasionally interrupted by the even more amiable presence of his pregnant wife, the filmmaker Caroline Lindy, the doc in its first half breaks down just what we talk about when we talk about AI. Is it narrow, general, super-artificial? Who are the field’s biggest players? It’s genuinely informative and mostly harmless material.
Even when there are contextual blind spots, such as the film’s too-brief look at the environmental effects of AI data centres, the information that is here is all packaged in a highly entertaining manner. Roher deploys cute stop-motion animation bits to break up the techno-babble, and layers all the info dumps atop a deeply personal, utterly relatable narrative. The director is concerned about how AI might destroy us all, yes, but mostly he’s worried about what kind of future his child is going to inherit. What parent wouldn’t do their best to pave the way for their children? Roher isn’t an engineer or a copyright lawyer, he’s a filmmaker. So a documentary it is.
Initially, his serious Terminator-coloured concerns of a robot uprising are echoed by a handful of AI skeptics and red-flag wavers who sit down for face-to-face interviews. The trouble begins, though, when Roher manages to convince a handful of the top AI entrepreneurs to make their case.
While Mark Zuckerberg and Elon Musk don’t end up in front of the camera, the filmmakers do get Google DeepMind’s Demis Hassabis, Anthropic’s sibling co-founders Dario and Daniela Amodei, and OpenAI’s Sam Altman on the record.
But while there are inquisitive pokes by Roher here and there, the interviews are simply too gentle, too safe, too frustratingly unproductive. When you have managed to secure time with Altman, arguably the leading candidate for reigning world supervillain, for an ostensibly carte-blanche conversation, then you have to put the screws to him. Especially if that person’s company has been waging a war against the film industry – against the very bones of the creative process – with relative impunity. (One recent good-news development that Roher wasn’t able to include here: Altman’s company this week suddenly shut down Sora, which allowed users to create short, AI-generated videos, and so spooked Hollywood that Disney signed on for fear of being left in the dust.)
AI startup uses Canadian casting networks to solicit actors, rattling industry
Indeed, for a film about the state of AI, there is a curious lack of attention paid to the future of, well, filmmaking. Or art of any kind. While there is one good bubble-bursting bit in which Roher doubles back on one expert’s rose-coloured contention that AI will allow his kid to pursue a creative life because AI will handle all the meaningless labour of our lives (economy? what economy?), I couldn’t help but watch the final act of The AI Doc with my brow furrowed and my stomach knotted. (Scanning the end credits, I was at least relieved to see not a single “AI” tool listed.)
As almost everyone on-screen contends, we are well past the point of return on AI – it is here, it is advancing more quickly than most think, so let’s learn to deal with it in the best way possible. But that means not letting twinkly-eyed proselytizers steamroll over us with their digital snake-oil slop, either.
Perhaps, though, I’m getting this all wrong, and The AI Doc is in fact designed to be a Rorschach test of sorts: you see your own internalized AI anxieties reflected back at you. Or maybe I’m being just as frustratingly have-it-both-ways as the movie itself and trying to justify what should be the most easily accepted truth of our time: There is no mechanical substitute for humanity. And there are no apocaloptimists in foxholes.