‘The AI Doc’ Review: Unproductive Both-Sides Documentary Is Too Little, Too Late
I suppose one day the Shop-Vac Corporation might install a TV screen inside a dust collector, but until then we’re going to have to face the facts: No movie exists in a vacuum. And boy howdy, is that true of Daniel Roher and Charlie Tyrell’s “The AI Doc: Or How I Became an Apocaloptimist,” a film about how large language models will impact the future of humanity, which can’t even keep up with how machine learning impacts us today.
The mass proliferation of what we sketchily call “artificial intelligence” is now an inescapable aspect of our daily lives, even though a whole lot of people want to escape it. And the news about AI is changing every day. “The AI Doc” was clearly produced before Elon Musk’s Grok started churning out non-consensual, sexually explicit images of real people, so it can’t speak to that.The film does mention that AI is an enormous drain on natural resources, including our water supply, but the United Nations only officially announced a week ago that we now live in a state of global water bankruptcy — and gosh darn it, that would have been relevant.
If Roher and Tyrell’s documentary were narrowly focused that might have been a quibble. But their “AI Doc” wants to be everything. A lot of people don’t understand LLMs very well, and they need a primer, so that’s the first part of the movie. But after they establish the basics, Roher and Tyrell set about the impossible task of cramming every point of view about AI into one, relatively short documentary. And that’s where they get into trouble.
The framework is that Roher, who won an Oscar for the 2022 documentary “Navalny,” is expecting his first child, and he’s afraid of what the future holds for his progeny. So he gets every talking head expert he can find — pro, con and mixed to the point of ambivalence — to visit his tiny studio and share their thoughts.
Roher asks everyone whether, considering what AI is doing to the world and our culture, and the potential, literal existential threat it poses to humanity, they would recommend if anybody had a child right now. The naysayers say nay. The CEOs of the AI companies, who have an extreme financial stake in convincing everyone to adopt AI as fast as possible, say that yes, you should absolutely have a child, because everything’s going to be great. “The AI Doc” lends equal credence to both sides, without really engaging with the fact that OpenAI CEO Sam Altman, and the head of Anthropic, and various other AI firm representatives have a gigantic conflict of interest.
Framing a documentary around the future of one specific child may sound like a smart way to keep “The AI Doc” focused and relatable. But the film keeps cutting back to Roher’s face, and leaning on his charged voice-overs, which has the unfortunate effect of making it seem all about him, specifically. This is about assuaging his anxieties, and he’s making sure the whole world sees what a concerned father he is. That doesn’t mean the interviews he’s cataloguing don’t have value, but it does lead the whole production to appear self-centered.
And that’s a huge problem because this story isn’t about him. You’d be hard-pressed to find a single person in the audience who isn’t currently affected by AI, or will be in the future. Tech CEOs spend a lot of the film saying that, thanks to their developments, we’ll soon live in a “post-scarcity” society, where nobody has to work, and you can do anything you want. What the movie only gets to later, and doesn’t delve into as deeply, is that a society where nobody has to work can’t function if it still runs on capitalism. The same people who brag that AI could replace your health insurance aren’t reckoning with the very simple reality that the existing health insurance companies still like making money, and will still want to get paid … by the countless of people who lost their jobs to AI.
I don’t envy Roher and Tyrell’s task. They’re trying to make “The AI Doc” the ultimate primer for the AI conversation, which could help a lot of people who haven’t thought about it much, or have only heard from one side of the argument. But in giving each side a long turn to speak, unchallenged, even the order in which the filmmakers choose to discuss these talking points can become a problem.
“The AI Doc” opens with desperation, then tries to instill us with hope, and then finally comes around to a generic “South Park” episode conclusion where, you know, Roher learned something today: both sides have problems and we all just have to do our best. That isn’t very productive, but since it concludes the documentary, it’s obviously intended to be our takeaway. Our unproductive, generic takeaway.
One last thought: “The AI Doc: Or How I Became an Apocaloptomist” debuted at the Sundance Film Festival in front of an audience of filmmakers and film lovers, yet it almost completely ignores the very real, immediate, practical threat to Roher and Tyrell’s own industry.
Their film acknowledges the danger of allowing governments to alter reality by manipulating videos and pictures to evil political ends. They even compare that to nuclear proliferation in the 20th century, although they don’t do a good job of addressing the fact that, even though the politicians claimed an arms race was an inevitability, it was fundamentally bad. You might think the most important lessons we learned from history would be relevant today, but I guess that wasn’t the point they were making.
The less life-threatening—but still livelihood-threatening—danger of taking art from artists, stealing their work, smooshing it together, and turning personal human experience into an averaged-out algorithm in which what frequently happens now always happens goes strangely unaddressed in “The AI Doc.” You would think the makers of documentaries—or any kind of film—would have an interest in that.
“The AI Doc” invites audiences to figure this all out for themselves, since we’re all experts in our fields, which is a very polite way to abdicate responsibility. So I guess that’s what I’m doing now, with this review. I’m figuring out that, whatever the intentions, a film like “The AI Doc” doesn’t do enough, go far enough or illuminate nearly enough about its subject. The pros don’t come from trustworthy sources and the cons require a lot more elaboration.
If anything it’s a little too much like AI. It’s the Cliff’s Notes version of a whole library, and I think we’d be better off going to the library.
The post ‘The AI Doc’ Review: Unproductive Both-Sides Documentary Is Too Little, Too Late appeared first on TheWrap.