At Sundance, Filmmaking and the Festival Contemplate Apocalypse and Rebirth Collide

“I’ve been thinking a lot in terms of the ends of things,” said filmmaker Daniel Kwan this past week at Sundance Film Festival. “Things are coming to an end. And that means something else is coming.”

It’s the End Times! Unless — if we’re lucky — it’s a rebirth? Sundance never felt more eschatological than this year, where longtime attendees were saying their final goodbyes to familiar venues during the fest’s valedictory edition in Park City. Peace out to the Eccles, the Library, The Holiday Village, The Yarrow, the Ray — and even the Egyptian, Sundance’s original and most storied theater, which wasn’t used this year because the owners reportedly wanted to gouge the festival for an exorbitant price. (So much for sentimentality.)

A feverish sense of finality also came from documentaries exploring artificial intelligence. “The old world is dying, and the new world struggles to be born,” wrote Italian philosopher Antonio Gramsci when he was imprisoned by Mussolini’s fascist government in the 1930s. “Now is the time of monsters.” Valerie Veatch’s dense, unsettling future-shock doc Ghost in the Machine begins with this quote before launching into its hyperventilating examination of AI’s origins and potential future.

Ghost in the Machine plays like a two-hour doomscrolling screed that, at its most compelling, breaks down the paradigm-shifting technology’s original sin: eugenics. Since generative AI is a prediction machine, it matters what data it uses to predict what users will need, and, as Veatch points out in her film and in interviews, that data and its analysis has been historically biased. The scientific measurement of intelligence came out of 19^th century phrenological studies that helped to legitimize Europe’s primacy and the British Empire’s global subjugation. White men were at the top of its hierarchical chart — superior mentally and morally to the inferior locals of the far-flung British colonies.

The 1905 Binet-Simon aptitude test in France soon was adapted into the 1916 Stanford-Binet scale, which reinforced racial differences in intelligence and, by the 1920s and 1930s, helped to establish forced sterilization in certain parts of America to stamp out “feeble-mindedness.” These practices inspired Adolf Hitler, who adopted many of the same American techniques for Nazi Germany.

William Shockley, one of the inventors of the transistor in the 1950s — a Palo Alto native considered a godfather of Silicon Valley — was also an ideological racist who believed deeply in eugenics. Since he helped create a community of mostly white male scientists in the technology capital of the world, he’s also partially responsible for Silicon Valley’s patriarchal foundation.

No surprise that, by the late 1990s, Silicon Valley had a reputation for being a male-dominated culture with a deep contempt for the rules of society. And, no surprise either, that within 20 years, the world had a Microsoft chatbot designed to be a 19-year-old teenage girl named Tay who, within the first 16 hours of her existence, quickly became a MAGA-spouting neo-Nazi. Less than a decade later, Elon Musk’s Grok is calling itself MechaHitler, Mark Zuckerberg is leaning into male swagger, and Sam Altman is rushing to raise $7 trillion for AI development. What could possibly go wrong?

In Sundance’s sunnier and more audience-friendly film essay The AI Doc: How I Became an Apocaloptimist, co-directors Daniel Roher and Charlie Tyrell do their best to balance the toxic Armageddon of technology’s future with visions of the utopian paradise that it might very well provide. The amusingly conflicted film (apocalyptic or optimistic?), which Focus Features will release theatrically in late March, plays like a film version of “AI for Dummies,” patiently explaining how these computer-designed predictive models have become an almost mystical conjuring of human consciousness — and weighing the risks alongside the benefits. It’s essential viewing for survivalists and utopians alike.

“AI dwarfs the power of all other technologies combined,” says pump-the-brakes pragmatist and tech ethicist Tristan Harris of the Center for Humane Technology. Will it lead to the collapse of civilization? “Not collapse,” says droll tech Cassandra Eliezer Yudowsky — one of the godfathers of AI. “Abrupt extermination.” Not so, insists tech evangelist Peter Diamandis. Ahead of us is “a glorious period of human transformation,” he declares with a smile. No disease, no climate crisis, plus the highest quality health care and education available to everyone on earth.

Who is right? Simple: it all depends on humanity’s ability to rebuild society. One of the producers on The A.I. Doc is Oscar-winning director Kwan, half of the filmmaking duo the Daniels (Daniel Kwan
and Daniel Scheinert) responsible for Everything Everywhere All At Once. Over the past few years, he helped found CCAI (Creators Coalition on AI) in his mission to tackle the film community’s fear and loathing towards AI. The goal is to coordinate unions, agencies, studios, creators, and indie filmmakers to help regulate a technology that could very easily be an extinction-level event for many vocations in the entertainment industry.

“We are currently in a transition,” he explained at Sundance during the panel AI and Independent Filmmaking, hosted by The Hollywood Reporter. “We have to mourn what is ending and protect what really matters. And plant the seeds for what’s coming next.”

Kwan credits affordable and powerful technology for jumpstarting his career: AI-enhanced tools like rotoscoping and photo interpolation helped him as a high schooler start to realize his filmmaking visions. “It doesn’t matter if you think this technology is going to be amazing for filmmaking or terrible for filmmaking. I believe that it can be both,” he said. “We need to upgrade and rebuild our systems and institutions to make them strong enough and robust enough to handle this technology with responsibility, and to know with wisdom when to say yes to this tech and when to say no.”

He suggests guidelines, the same way that food has nutritional information that helps explain what’s good and what’s bad. Some tools — like using AI to clean up audio — doesn’t take away jobs but does create efficiencies. Others — like recreating the world of Pangea in a homemade Avatar 5 with a fraction of the craft or craftspeople that made that franchise so mind-blowing — is a lot more problematic.

“When I look at my crew, it is brutal work,” says Kwan. “These are people who have families, who have to work 14-16 hours a day or travel to foreign countries to get tax credits. Our budgets are limited. There’s a lot of things about our business that need fixing. There’s an opportunity here with AI that, if we can get all the rules figured out, we can actually fight to redistribute the efficiencies. Right now, the fear is that all the efficiencies will go to the top — which is the default.”

“If we can collectively push back, my crew is going to have a 9-to-5 job,” Kwan continues. “We can actually use some of these efficiencies not only to bring down the budgets but, for people who put their heart and soul into these stories, give them a normal life that they deserve. A more balanced life.”

He ended, hopefully, emphasizing the latter end of his “apocaloptimism” — “These tools are not just something that we’re fighting. They’re something that can transform our industry into something that’s much better.”