For over a year, our insiders have been telling us the same thing: Hollywood is far deeper into AI than the studios will admit out loud.
Disney is restructuring around it.
The incoming bosses of the merged Paramount-Warner Bros. are planning for it.
Wall Street is demanding it.
Now comes confirmation of just how organized it all is, as the most powerful people in the industry have been meeting about AI in secret since November 2023.
The Secret AI Talks Go Public
The Ankler‘s Richard Rushfield, who sat in on the meetings under an off-the-record agreement, revealed on August 20 that legendary producer Kathleen Kennedy — the longtime Lucasfilm boss — and American Film Institute dean Susan Ruskin have been quietly convening Oscar-winning filmmakers, big-name actors, studio C-suite executives, and Silicon Valley players for private AI conversations.
The first session happened in November 2023 at the Los Angeles studio of AI artist Refik Anadol, in the shadow of the strikes. The talks then moved to the American Film Institute campus, and in nearly three years, not a word leaked.
Last weekend, the group expanded to roughly 120 participants and produced its first public document, titled “Human Generative Workflows,” which The Ankler published in full.
Jon Avnet, who co-chairs the DGA’s AI committee alongside Christopher Nolan, led one of the panels and said the point was to get the conversation happening “in light of day.”
Bryn Mooser, founder of artist-led generative AI film and animation studio Asteria, put it more bluntly: using AI is a filmmaker’s choice, but understanding it isn’t optional.
The Line Gets Drawn Where The Money Is
The document’s core argument is that one word is doing too much work.
“AI” currently covers everything from noise-reduction features that have been sitting inside editing software for years to a deepfake performance made without consent, and the group wants that split into three categories for final footage.
Utility and embedded AI
Covers the machine-learning tools already built into Premiere, Avid, and DaVinci Resolve: denoising, upscaling, rotoscoping, wire removal. Copyrightable.
Human Generative Workflows (HGW)
Covers generative models working inside professional pipelines like Nuke, Unreal, and After Effects, under artist control and often trained on the team’s own artwork. The group argues this is simply the next step after CGI: human-authored and copyrightable.
Machine Generative
Pure prompt output: synthetic performances, voices, and stories that replace human creative work. Not copyrightable and, per one producer in the room, where the word “slop” comes from.
Here’s why “copyrightable” is the whole ballgame
If a studio can’t copyright footage, it doesn’t own it, meaning it can’t license it, sell it, or stop anyone from copying it.
The U.S. Copyright Office has said typing prompts alone doesn’t count as human authorship, and nobody is building a $200 million tentpole on footage they can’t own.
So the group’s map draws the legal fence around everything short of pure prompting, which happens to protect exactly the kind of AI use the studios are already ramping up.
Dear Upstairs Neighbors
The document’s showcase is Dear Upstairs Neighbors, a six-minute animated short that premiered at Tribeca in June.
Every frame was generated by fine-tuned AI models, made by a team of about 45 people, including Pixar and DreamWorks veterans working alongside Google DeepMind engineers .
The group argues that every frame is entirely human-authored.
It also notes the credits war is already underway: A24’s Heretic declared no generative AI was used, Vince Gilligan’s Pluribus signed off with “This show was made by humans,” and Bad Guys 2 used its credits to ban AI training on the film.
Everything Our Insiders Described Is Already In Motion
Back in July, we exclusively reported Disney is restructuring around AI to push out more content with fewer people, the same stretch that saw Marvel eliminate its visual development team and lay off legendary artist Andy Park.
Our sources also detailed the plan for the merged Paramount-Warner Bros.: Mike De Luca and Pam Abdy organizing the studio with in-house and outside production and post-production teams built to maximize AI on tentpole films.
On August 11, activist investor Anson Funds told Lionsgate to pivot to AI or sell before it becomes an “AI casualty,” language that lined up with our Disney reporting almost word for word.
And it’s already on screen.
Supergirl used AI to design Lobo, Warner Bros. pulled the featurette that revealed it, and our insiders say James Gunn is using AI on Man of Tomorrow.
The official Spider-Man: Brand New Day art book shipped with apparent AI concept art, the same artificial look our review flagged in the finished film.
Disney, meanwhile, signed a landmark deal putting Marvel, Star Wars, and Pixar on OpenAI’s Sora, which has since been canceled due to OpenAI abruptly shutting down its AI video platform, Sora, over costs.
Marvel even used it for the Secret Invasion intro, and the rights owner of Stan Lee is “bringing him back” through use of AI.
Same Sources, Same Track Record
The group that held the meetings with Kennedy frames all of this as keeping AI under human control, a line drawn in favor of artists.
But look at what the document actually does: it takes most AI use in production off the “controversy” table and stamps it as legitimate, ownable, and normal. And, more importantly, copyrightable and sellable.
While the industry publicly wrung its hands, its most powerful people spent nearly three years in closed rooms working out how to live with the technology, and the first thing they published is the permission structure.
That is the Hollywood our insiders have been describing all along: Disney restructured for it, the next Warner Bros. regime is being staffed for it, and per those same sources, the machines are already in the pipeline on Man of Tomorrow and more.
The meetings are public now. Expect the AI to follow.
