GENERATIVE AIBen Affleck Says Netflix's 'Animals' Used AI Post Tools Trained on a Proprietary Dataset
VP LandOct 26, 2026 · 3 min read
Ben Affleck confirmed that his Netflix feature Animals used AI tools in post-production trained on a proprietary dataset, which he said was created so it "wouldn't have to be extractive," according to The Hollywood Reporter. He described the technology as a discrete set of post tools that filmmakers at Netflix can choose to try.
- Training data stayed proprietary. Affleck said he won't use AI that "trains on or has learned from any work that is copywritten," which is why "we created a dataset that was proprietary."
- Affleck calls it a discrete post tool. He described "a discrete set of tools" and "another interesting post tool," adding, "It's not about making fake people or replacing anybody, ever."
- Invisibility is the quality bar. Affleck said that "if you can see it or if you can tell what it is," the result is ruined, the same way a bad visual effect or a bad performance fails.
- The specific uses remain undisclosed. At Bloomberg's Screentime event, he said he used AI "for lots of things in that movie" but was hesitant to point out which ones.
- Netflix owns his AI company. Affleck sold InterPositive to Netflix for nearly $600 million, THR reported.
Affleck says Animals avoided AI trained on copyrighted work or peers' work
Affleck directed, co-wrote and stars in Animals, a film about a Los Angeles mayoral candidate whose son is kidnapped, with Kerry Washington playing the candidate's wife. Animals hits select theaters and starts streaming on Netflix on October 9.
He told THR his approach to AI is "really simple." He won't use a system that learned from copyrighted work, "or that is an intelligence created by a peer of mine or somebody that I look up to."
"That was why we created a dataset that was proprietary, so that it wouldn't have to be extractive," Affleck said.
He said he has no interest in making movies where you "type things and things come out," and doubts audiences would care enough about that kind of output to want to see it.
We have previously questioned InterPositive's post-training approach. The THR interview does not say whether InterPositive's tools were used on Animals.
Affleck describes Netflix asking filmmakers whether they want to try the tools
Affleck described the tools as something filmmakers at Netflix are invited to try. "And like at Netflix, it's like, 'Hey, do you want to use this? Do you want to try this?'" he said.
His stated thesis is that giving the technology to people who make films produces better results. "You're going to get way better results. And we have, and we did," Affleck said.
Netflix used generative AI in roughly 300 titles, with that use concentrated in post. Affleck did not say whether Animals is part of that count.
Affleck wants AI inside guild-signatory companies with clear compensation for artists
Affleck said his goal is to bring the technology into "these companies who are guild signatories, who have relationships with the creative community." He contrasted them with platforms that might "put you in a movie without your permission and without paying you."
He said he is "already uncomfortable with some of the ways some tech companies behave around image and likeness." He argued the industry should not abstain from the technology and let others take it and dominate it.
Affleck also said AI is "deeply and fundamentally limited in a lot of ways." He pointed to theatrical attendance, labor stoppages and production tax rebates as Hollywood's real problems.
He pointed to Artists Equity, his company with Matt Damon, and its belief that "in success, everybody is compensated in a very clear way."
Affleck said InterPositive, which Netflix acquired to bring its AI post tools in-house, shares that philosophy. In his description, a director or actor who contributes their own art can choose to train on it and use that intelligence.
"You don't have to give it to the rest of the world. They are not entitled to that," Affleck said. He added that filmmakers can use the technology "if you feel that it helps your film in one way or another, and there are places where it does."