The AI Doc review: a hype piece for doomers and accelerationists alike
We’re in the midst of a massive campaign to integrate generative AI into almost every aspect of our lives, but it’s still easy to be confused about what it is and how it works. It doesn’t help that many supporters and detractors of the AI generation both talk about it with feverish hyperbole that sounds like fantasy advertising copy. And the speed at which AI companies release new iterations of their products can make it difficult to keep up with what’s happening across the industry.
In The AI Doc: or how I became an apocaloptimistco-directors Daniel Roher and Charlie Tyrell attempt to make sense of this moment in the rise of the AI generation. The film features researchers, developers, and CEOs of AI-generation companies – the exact people you’d want to see a documentarian talk to about this genesis and possible future of this technology. But as much excellent access as The AI documentThe production team managed to get it right, but the documentary barely makes an effort to use it effectively. The AI document is imaginatively produced with clever art direction, but it lacks substance and says nothing truly insightful about its subject. At a time when people could really use a thoughtful introduction to how Generation AI is already impacting their lives, this documentary fails to meet the moment.
The AI document is also the story of one man’s (co-director Roher) general concerns about the impacts of the AI generation on society. At the start of the film, Roher (who won an Oscar in 2023 for his documentary Navalny) comes off as someone who doesn’t have the best understanding of what models like ChatGPT, Claude, and Gemini actually are. But he’s heard alarming headlines about how the AI generation could give rise to feeling machines that destroy humanity, which scares him because he and his wife – Caroline Lindy – are expecting a baby. Roher wants to better understand this new technology which leads him to wonder what kind of world his child will be born into. So he sets out to speak with a number of experts with different perspectives on AI.
The documentary is structured into four acts that trace the arc of Roher’s feelings as he interviews AI doomers, accelerationists, academics, and some of the industry’s most powerful executives. Roher leads with pessimists like Center for Humane Technology co-founders Tristan Harris and Aza Raskin, both of whom present AI as an existential threat that could lead to societal collapse. An interviewer insists that there could be a robot uprising that results in the destruction of humanity, and the documentary cuts to clips of The terminator And The Matrix. And in response to Roher’s question whether an apocalypse scenario is on the horizon, the documentary’s AI critics often respond with worrying variations of “maybe” and “probably.” This type of doom and gloom and doom talk is one of the most prevalent forms of advertising that AI companies have used to convince people that their products should be taken seriously.
Roher – who presents himself as a sort of naive audience surrogate – seems to take these statements literally – particularly in moments where he turns the camera on himself to emote about the gravity of his impending fatherhood. Notably, The AI document never takes the time to explore how AI has disrupted aspects of filmmaking, which might concern an artist/director like Roher, whose hand-drawn sketches and paintings are used throughout the documentary to visualize his feelings. The lack of commentary on the impact of AI on Hollywood and the lives of creative professionals is particularly egregious because of the scale The AI document relies on animated sequences produced by Toronto studio Stop Motion Department to illustrate its subtleties.
Roher’s gloomy outlook on AI begins to change as the documentary introduces optimists like Anthropic president/co-founder Daniela Amodei and LinkedIn co-founder Reid Hoffman, who insist that it’s actually a great time to become a parent, because AI is poised to open up all kinds of new possibilities in a future utopian society, like tailored, easily accessible healthcare. It feels like Roher is trying to give the public the “fair” view by contrasting these two sides of the AI debate. But by giving its doom and gloom voices so much time to present AI’s most hyperbolic potential outcomes with little hesitation, the first half of the documentary feels more like an overlong commercial for the technology than a measured analysis.
Image: Focus characteristics
The AI document is on much firmer footing as it pivots toward conversations with journalists, including Karen Hao and whistleblowers like Daniel Kokotajlo, who talk at length about how AI products are a reflection of the companies that build them. While the first two segments of the film present generative AI as an almost magical thing that cannot be fully understood, the third explains how LLMs are really just sophisticated pattern recognition machines that must be trained on huge amounts of data to work. The third act also briefly touches on some of the real damage currently being done by the strong push for AI. But because The AI document moves through each of its segments so quickly that some of its more pointed observations — like how these companies rely on brutal, underpaid human labor to process their data sets — about AI don’t receive as much prominence as they should.
At some point, Roher recognizes that every conversation he has will feel stale by the time The AI document versions due to the speed with which AI is advancing and being deployed. This becomes especially true when he speaks with Sam Altman, director of OpenAI, and Dario Amodei, CEO of Anthropic. Roher had no way of knowing his film would debut at a time when Altman is under fire for striking a deal with the Department of Defense to provide models that could be used for mass domestic surveillance. Roher also couldn’t predict that Amodei would spend weeks fighting with the Pentagon over Anthropic’s refusal to give the government unfettered access to its technology (and that its AI would be used to strike Iran). But when one approaches the film with some awareness of what’s happening in the news, Roher’s softball questions to these industry leaders about their feelings about the future seem superficial.
As companies and governments continue to introduce AI into virtually everything, the public needs more thoughtful questions about the technology that allow them to fully understand its potential benefits and the ways it can be used against them. Unfortunately, The AI document is not up to par.
The AI Doc: or how I became an apocaloptimist hits theaters on March 27.
