Hollywood remains fascinated by artificial intelligence as a subject. Whether it should become part of the creative process is proving far more contentious.
Guillermo del Toro used San Diego Comic-Con to draw a clear line around the 3D conversion of Pan’s Labyrinth.
“Obviously I said, ‘Absolutely no goddamn AI,'” del Toro told the Hall H audience.
The contrast was difficult to miss. Some of Comic-Con’s most prominent genre announcements, including Neuromancer, Blade Runner 2099, Dark Matter and Lanterns, examined artificial intelligence, surveillance, hacking and human-machine relationships. Hollywood remains comfortable using AI as a dramatic subject. It is far less settled about using it to make the work.
Guillermo Del Toro Rejects AI In ‘Pan’s Labyrinth’ 3D Conversion
Del Toro appeared on Friday for a panel marking the 20th anniversary of Pan’s Labyrinth, alongside Doug Jones, Ivana Baquero and cinematographer Guillermo Navarro. The film returns to theaters October 9 in 4K, 3D and HDR. The conversion was completed by hand.
His concern extends beyond protecting individual positions. What is being defended, he said, is a lineage of art. Cutting one generation off from learning its craft also separates the medium from the knowledge accumulated before it.
Christopher Nolan has expressed a related, although slightly more qualified, position. He accepts that AI may provide useful imaging tools but has called the idea that it could replace human creativity wholesale “nonsense.” He has also argued that the technology, and the motives of those promoting it, should be approached with skepticism.
Both directors leave room for technology to assist production. Their objection is to treating it as a substitute for human judgment, training and responsibility.
It is tempting to frame the debate as a contest between established directors defending craft and technology companies promising cheaper production.
My recent interviews with more than 10 AI filmmakers suggest the divide is more complicated.
AI Filmmakers Say Human Craft Still Shapes The Finished Film
Hal Watmough, a BAFTA-winning editor who directed the predominantly AI-made short The Trials, began from a position of anxiety.
“I was kind of concerned about this thing that was going to take our jobs,” he says. His response was to make the film as “an invitation and not a threat” and to show other filmmakers what remained possible.
The 33-minute production involved writing, editing, concept art, voice direction, grading, sound design and repeated rounds of rejection and refinement. AI generated material inside that process; it did not remove the need for it.
Dave Clark, cofounder and chief creative officer of AI entertainment studio Promise, makes a similar distinction between generating a striking clip and completing a film.
“You can’t prompt your way to the ends,” Clark says. “You can prompt your way to some of the things that you need, but then you have to use all your filmmaking knowledge to put everything together.”
Promise brings in former Oscar-winning visual-effects supervisors for precisely that reason: people who can absorb a producer’s notes at 1 a.m. and rebuild work from scratch.
“That’s not just a prompting exercise,” Clark says. “That’s a filmmaking exercise.”
Nik Kleverov, director and co-producer of the AI-assisted feature Critterz, offers perhaps the clearest measure of success.
“When people watch the movie, I just want them to watch a great film and not think about AI once,” he says.
Zack London, the YouTube creator known as Gossip Goblin, spent three years building an audience across hundreds of short films before attempting a feature, and argues that the world and the story have to come first.
“AI is just a practical tool to execute on that,” London says.
His anthology feature Gods Don’t Give Gifts reaches U.S. theaters on October 30. It uses voice actors, musicians and a Foley artist rather than generating those elements, and London treats performance, sound and writing as decisions he does not want to delegate.
Wonder Studios Rejects The One-Person AI Film Studio
Wonder Studios cofounder Justin Hackney also rejects the image of the future filmmaker as one person producing an entire film alone at a keyboard.
“I just don’t think… all the filmmakers of the future will also be all the people… pressing the button,” he says.
As productions grow, he expects collaborative groups built around smaller crews, faster pipelines and new combinations of traditional and technical skills.
His cofounder Xavier Collins is more candid about the employment consequences.
“There are cost implications, timing implications, and job implications,” Collins says.
Some work will be replaced or performed differently. But he argues the result will not be simple job destruction: “I think it will be job replacement, job changing, and job creation.”
That does not answer del Toro’s central concern. Entry-level work is how industries reproduce their craft. Remove too much of it and Hollywood may save time and money while weakening the route through which future directors, editors and artists learn.
The practical divide is therefore not simply between filmmakers who use AI and those who refuse it. It is between production models that use the technology to extend human craft and those that use it to remove the people who practice it.
Comic-Con’s contrast was not only cultural. Hollywood is happy to tell stories about AI. Its harder test is whether it can use AI without breaking the creative lineage del Toro wants to protect.