Gal Gadot went on Josh Horowitz’s Happy Sad Confused podcast to promote an action film and ended up describing a contract negotiation instead. The film in question is Doug Liman’s Bitcoin, and the negotiation ran for half a year.
“This was the longest contract I have ever worked on,” she said. “It was six months from my lawyers to go through every scenario we could think about as far as the AI component of it. I did not want AI to have anything to do with my performance. We had to protect that, to such an extent that SAG called my lawyers to ask for pointers to update the SAG contract to protect other actors. As far as the performances, it’s all me.”
That last detail is the story. A union representing roughly 160,000 performers called an individual actor’s legal team to find out how they had written a clause. Not the other way round.
The short version
- The film: Bitcoin, directed by Doug Liman, starring Gadot with Casey Affleck, Pete Davidson and Isla Fisher. Shot over roughly 20 days on a largely bare soundstage
- Where the AI sits: production design and lighting. Sets, environments and the look of the world are generated rather than built
- Where Gadot says it does not sit: the performances. “As far as the performances, it’s all me”
- The money: producer Ryan Kavanaugh has claimed the $70 million production saved north of $200 million against what a conventional shoot would have cost
- Her actual warning: not that actors should refuse the work. That they should assume the standard contract does not protect them and negotiate it line by line, as she did
- The part nobody negotiated: the clause covers her likeness and her performance. The savings come from the departments that had no seat at that table
What she actually secured
Strip out the noise and Gadot bought herself one specific thing: a guarantee that no model touches her performance after the fact.
That is narrower than it sounds and more important than it sounds. The risk in a generative production is not primarily that a studio replaces an actor wholesale. It is the small, deniable adjustment made in post: a mouth reshaped to fit a redubbed line, an expression nudged warmer, a body movement smoothed. Each edit is trivial in isolation. Collectively they mean the performance the audience sees was finished by software after the actor went home.
The concern was not hypothetical on this production. The film’s UK casting notice reportedly reserved the right to “change, add to, take from, translate, reformat or reprocess” performances using generative AI and machine learning, language broad enough to cover exactly those lip, facial and body adjustments. That notice is what turned a budget story into a labor story.
Six months of lawyers bought a carve out from that. And the reason SAG-AFTRA phoned afterward is that its own agreements were written for a world where the fight was over digital replicas and full body scans, not over whether a real performance can be quietly edited by a model.
The savings did not come from the cast
Here is the part that sits uncomfortably alongside the good news, and it is worth stating plainly rather than sharply.
If the producer’s numbers are anywhere near right, a very large sum was removed from this production’s budget. That money did not come out of the leads’ fees. It came out of the departments that build things: construction, set decoration, lighting, locations, the practical crafts that fill a frame. Those are the jobs the generated environments replace.
Gadot’s own framing acknowledges the shape of it. “The AI component in this movie is for sets and the lighting. We had a broad crew and a lot of people were working on this movie.” Both halves of that are probably true. A 20 day soundstage shoot still needs a substantial crew. It also needs far fewer people than a production that has to physically construct every location it shows.
“The train has left the station”
Gadot’s defense of taking the job is the most quoted part of the interview, and it is more honest than the usual industry line.
“I am afraid of AI, but I am going to be afraid and look away? No,” she said. “I am afraid but I am playing with it and learning it and trying to educate myself. The train has left the station. You’re going to work with it or be out of the game completely. I am trying to make the best of it and protect what I can protect.”
Read the whole quote and the criticism aimed at her looks slightly misdirected. She is not claiming the technology is good for the industry. She is saying it is arriving regardless, and that the useful response is to arrive with lawyers rather than with a statement.
Whether that is courage or capitulation depends on your view of how much leverage any individual has, and reasonable people land in different places. What is not in dispute is the outcome: her clause exists, it works, and the union is now studying it.
| Element of the production | How it was made | Contested? |
|---|---|---|
| Performances | Filmed conventionally. Gadot says no model touches them | No, and only because it was negotiated for six months |
| Sets and environments | Generated. Actors shot against a largely bare stage | Yes. This is the labor argument |
| Lighting | Substantially generated in post rather than rigged on set | Yes. A large, skilled department |
| Shooting schedule | Around 20 days on one custom soundstage | Indirectly. Short schedules mean fewer crew weeks |
The subject is doing a lot of unintentional work
It is hard to ignore what this film is actually about.
The story circles Satoshi Nakamoto, the pseudonymous creator of Bitcoin whose identity has never been established. Since you cannot cast a person nobody has identified, the narrative reportedly centers instead on Craig Wright’s public claim to be Nakamoto, played by Casey Affleck, with Pete Davidson as the billionaire miner Calvin Ayre.
So: a film assembled by generative models, about a man who said he was someone he was not, concerning a currency whose inventor may not exist as a single person. The Bitcoin story has generated stranger artifacts, including a video game about digging through a landfill for a discarded hard drive, but this one has a tidiness that no screenwriter would risk.
The honest read
Gadot’s warning to other actors is real, and it is narrower than the headlines suggest. It is not “refuse these films.” It is closer to: the paperwork you are handed does not protect you, protection has to be built clause by clause, and it took a major star’s legal team half a year to build it once.
Most working actors do not have six months of legal budget. That is the actual problem, and it is why SAG-AFTRA calling those lawyers matters more than anything Gadot said about her own comfort level. A protection that only exists for people who can afford to invent it is not a protection, it is a privilege, and turning it into a standard term is the only thing that changes the math for everyone else.
The wider pattern is familiar by now. Generative systems tend to arrive first for the work that is invisible in the finished product, which is precisely the work whose absence nobody in the audience will notice. Amazon effectively conceded the point when it shut down the marketplace it had built for the human labor behind automated systems, and the same substitution has already reshaped the open web, where roughly a third of pages created since ChatGPT launched were not written by a person.
A film with real actors on generated sets is that same trade, staged where everyone can see it. Gadot protected her half. The other half was never hers to protect.
