It is bold of Daniel Roher and Charlie Tyrell to call their film The A.I. Doc, as though it is the definitive statement in the documentary space on the subject of A.I. when it is merely one of many that have been produced and no doubt many more to come. But those other docs may not have the backing of the makers of Everything Everywhere All at Once or the legitimacy (earned or not) of some of the biggest figures in the A.I. space being involved. These are the kind of things that can give it a profile, and a theatrical run so rare for a documentary (First Lady vanity project aside). And it is worth being thankful for the privileges bestowed on this film, because its subject matter is incredibly important. Which is also why it is a fortunate thing the movie handles that subject matter responsibly.
The most critical thing is that for a movie about A.I., Roher and Tyrell ensured that it’s focal point was entirely human. Roher is a successful documentarian (his film Navalny won the Oscar a few years back), his wife Caroline Lindy is likewise a filmmaker in her own right. Both are tuned in to the buzz around artificial intelligence and its consequences for the future, which they find themselves having particular anxiety around when they realize they are expecting a baby. Confronted with a series of fears and difficult questions about the world his child will inherit, Roher conducts a series of interviews with a variety of experts, researchers, and insiders to better learn about A.I., and what it really means for our collective future.
Broadly, the film is split across four acts -not including its introduction- painting a variety of images with appropriate visual accompaniment in little doodle and stop-motion animations of the world artificial intelligence is believed to be ushering in. Perhaps smartly, it rips the band-aid off early, pairing some basic definitions of what A.I. is (with a point made of how little consensus there is on that) with observations and extrapolations in the grimmest directions imaginable. This first collection of interviewees comprises many researchers and activists in the field of A.I. Risk, including Tristan Harris, Dan Hendrycks, Aza Raskin, Connor Leahy, and Ajeya Cotra, scholars like Yoshua Bengio and Emily Bender, and even author Yuval Noah Harari, each discussing with clarity the gravity of A.I. advancement on the survival of human society. Points are made about the ethical ambivalence of A.I., the lack of control over its self-development especially where large-scale weapons are concerned, and the scale of its intelligence easily dwarfing our own. A.G.I. (Artificial General Intelligence) is discussed as the inevitable next step, which would fully render the input of humans in all manner of fields and disciplines obsolete in vivid and cataclysmic ways. It is sobering and broadly pessimistic -doomerist even. Harris shares how some in his field don't expect their children to graduate high school, Eliezer Yudkowsky posits an "abrupt extermination" -perhaps within the next decade- rather than a gradual collapse. Dark and deeply depressing stuff.
Roher naturally is driven to counteract with a series of more positive testimonies from a different collection of A.I. enthusiasts and entrepreneurs, whose takes on the emerging technology and its effects for humanity are not only positive but downright Utopian. Figures like Peter Diamandis, Daniela Amodei, Guillaume Verdon, and Reid Hoffman emphasize the apparent limitless promise of A.I.; the scientific and labour advancements it is expected to usher in -thereby granting people more personal freedoms, a symbiotic relationship of mutual growth and intelligence, and that rather than being fearful, Roher should be happy to have a child at such an exciting time in human history. Though for as catastrophically gloomy as some of the prior comments had been, most of these grandiose predictions feel just as over-indulgent -especially when noting a plurality of the talking heads are effectively in the business of salesmanship. Still, the film gives credence to their perspective and doesn't diminish the allure of what they are saying.
But this vision too is not allowed to go unchecked. Each chapter is nicely broken up by Roher talking to Lindy, who may be the most truly rational voice of the whole film -pushing him on to find some answer to the AI.. question that is not one of two extremes, for the sake of their child. It is the third chapter that is the most important, which returns broadly to the space of the skeptics but redefines the stakes as being about powerful individuals and institutions rather than the technology itself. As journalist Karen Hao points out, A.I. is not some mystical thing that dropped out of the sky -it is a tangible technology that is engineered and developed by people; people who need to be held accountable for it. Sure enough, it culminates in Roher sitting down with three of the five most powerful CEOs in the A.I. space -Sam Altman (of OpenAI), Dario Amodei (of Anthropic), and Demis Hassabis (of DeepMind) -Mark Zuckerberg declined to participate and Elon Musk changed his mind at the last minute.
These figures are put on the stand effectively, but there is no dramatic crossfire or confrontation -the movie isn't about that, and Roher comes at it from a very relatable place (he and Altman even bond a little over their mutual impending fatherhood). In this context, their responses, predictions (and deflections) related to A.I. safety and regulation mean much more. Without being direct, the movie is actually very clear on what is going on at a structural level -the economic and geopolitical boogeymen being exploited for the proliferation of this technology and the profits for its corporate drivers. “You can’t separate the promise of A.I. from the peril,” Harris says at one point. The stress of responsible guardrails is the movie’s argument, which is made fairly well. Roher comes off as a very neutral party. He gives each argument its moment and takes them seriously -perhaps too much so at some junctures when it comes to the pie-in-the-sky crowd- but his conclusions are not relegated to a wishy-washy both-sidesism, if he does perhaps play down a little bit of the chaos for the sake of the movie’s tone. He comes close to a kind of saccharine note to end on for hope for the future, but pulls back sufficiently before the movie’s actual resolution. And it is not one that finds an answer, as any reasonable audience could surmise going into the film. But arguably one of the most important goals of a documentary -especially on an “issue”- is to raise awareness, and that is what this film does. It may well be the 2020s equivalent of An Inconvenient Truth, though with a little bit more charm between its reams of anxiety.
Respectably, there is no apparent A.I. in the construction of this movie, apart from footage extracted from other sources. Though it is something of a glaring omission that the movie doesn't discuss at any point in depth generative A.I. and its effect on the arts -especially considering Roher is himself a filmmaker. It might not seem as critical as the variants of A.I. that take up a larger focus here but it is very important, and just as necessary to be wary of and regulate. Human ingenuity is on display though in the frequent stop-motion animation used as a compelling visual aide, illustrating well the breakdowns of A.I. behaviours and functions, as well as the metaphor of Roher climbing a mountain. The music and editing are really on point as well, enhancing and punctuating the beats of Roher's responses to his interviewees, and creating harrowing or beautiful collages of new, preexisting, or personal footage.
"Apocaloptimist" is a term dropped late in the film by National Security expert Jason Matheny as a median role between the doomers and the dreamers, anticipating a lot of bad things but hopeful in the capacity of people -both as individuals and collectives- to pull through and weather the A.I. storm. It is where Roher (and presumably Tyrell) is at by the end of this enterprise, and the audience is encouraged to come to that space as well. It may not work, but The A.I. Doc is a valiant effort regardless. It is important and revelatory, whatever its attitude may be. If artificial intelligence is indeed not going anywhere, movies like this are vital for our understanding of the technology and for charting our way through its hopeful or terrifying brave new world.
Support me on Patreon: https://www.patreon.com/JordanBosch
Follow me on Bluesky: https://bsky.app/profile/jordanbosch.bsky.social
Letterboxd: https://letterboxd.com/jbosch
Twitter: https://twitter.com/Jordan_D_Bosch
Comments
Post a Comment