Google Is Offering Hollywood Hundreds of Millions to Train Its AI on Movies and TV Shows
The tech giant wants studio libraries. The studios want the money. But the workers and audiences caught in the middle may pay the real price.

Key points
- Google has approached Disney, Warner Bros. Discovery, and Universal about licensing deals to train its AI models on copyrighted film and TV content, according to The Los Angeles Times.
- Google could pay a single studio around $40 million for the rights to generate AI images of one copyrighted character, with total deals potentially reaching into the billions.
- No major studio has signed an agreement yet, though Google's DeepMind research division struck a $75 million investment deal with film studio A24 earlier in 2025.
- Lionsgate signed a licensing deal with AI video company Runway in 2024, but nothing finished has been released publicly.
- Fans reacted negatively when AI-generated images appeared in the official artbook for Spider-Man: Brand New Day, a film that still earned $2 billion at the box office.
Google is quietly knocking on Hollywood's door, chequebook in hand. The company has been pitching major film and television studios on licensing agreements that would let it train its AI systems on their back-catalogues of copyrighted movies and characters, in exchange for enormous sums of money.
No deal has been signed. Disney, Warner Bros. Discovery, and Universal haven't agreed to anything, according to unnamed industry sources cited by The Los Angeles Times.
What is Google actually offering?
The numbers are striking. Google could pay a studio roughly $40 million for the rights to generate AI images of a single copyrighted character. The more characters a deal covers, the higher the total climbs, potentially into the billions. Studios might also receive a cut of any advertising revenue Google earns from AI-created content featuring their intellectual property on platforms like YouTube.
The goal, from Google's side, is straightforward. If its Gemini AI system can produce officially licensed images of beloved characters, ordinary people start seeing AI as something fun and familiar rather than threatening. That kind of goodwill is hard to buy any other way.
Google's DeepMind division, its main AI research arm, already took a step in this direction earlier in 2025, putting $75 million into A24, the studio behind Everything Everywhere All at Once. We've tracked DeepMind's moves since 16 July 2026; this investment is the most commercially expansive yet.
Why haven't the studios said yes yet?
The risks for studios are real, even if the paycheques are tempting.
Fans were genuinely upset when AI-generated images turned up in the artbook tied to Spider-Man: Brand New Day. The backlash arrived after the film was already in cinemas. Nobody knows whether earlier disclosure would've dented its $2 billion box office total, but studios are clearly watching the reaction carefully.
There's also the workforce question. AI's appeal to studios is partly that it could speed up production and cut costs. Shareholders would welcome that. But making more films more cheaply would likely cost a significant portion of industry workers their livelihoods. That tension doesn't disappear just because a licensing deal is framed as a creative partnership. It's a dynamic our earlier story captured well: award-winning screenwriters are already taking gig work teaching AI systems their craft, earning as little as $12 an hour to do it.
Separately, on 31 August 2026 we reported that Sony and Warner are suing Anthropic over unlicensed training data, which tells you something about the legal temperature in the room when Google's executives arrive with their pitch.
Lionsgate signed a licensing agreement with AI video company Runway in 2024. The two companies say they're working on AI-generated short films, but nothing has been released.
What does this mean for audiences?
If one big legacy studio signs with Google and releases something, others will feel pressure to follow. The first to go public with a major AI-made project will find out quickly how audiences respond when AI involvement is disclosed upfront rather than discovered later.
For now, the studios appear to be waiting to see who blinks first. Google has every reason to keep pushing. Access to decades of beloved film and television content would give its AI a quality and cultural legitimacy that unlicensed training data simply can't match. The money on the table is real. So is the risk that audiences feel the entertainment industry sold something they loved to the highest bidder.
Common questions
Does this affect films I watch right now?
Not yet. No major studio has signed a deal, so nothing trained on licensed Hollywood content has reached a cinema screen.
Would studios have to label AI-generated content?
No legal rule currently forces studios to disclose AI use in most markets, though the negative reaction to the Spider-Man artbook suggests audiences will notice and react when they find out.



