Skip to content
London · Since 1999
Cinema

Hollywood’s Stars Make Peace With AI and Put Their Faces Up for Sale

Cate Blanchett, Madonna and Hollywood's stars want AI to ask first and pay. Voice actors have priced their digital doubles. No machine can copy an actor's emotional memory.

Three-panel montage: Cate Blanchett smiling on the Venice red carpet, Michael Caine seated in glasses before a red curtain, and Madonna singing into a microphone in a black hat and coat.
From left: Cate Blanchett (Photo: Harald Krichel / Wikimedia Commons, CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0); Michael Caine (Photo: Manfred Werner - Tsui / Wikimedia Commons, CC BY-SA 3.0, https://creativecommons.org/licenses/by-sa/3.0); Madonna (Photo: Voice of America / Wikimedia Commons, public domain)

On Friday Hollywood’s stars marched into the biggest copyright battle of the AI age carrying an invoice. Cate Blanchett’s Human Consent Foundation, the talent agency CAA and groups representing musicians and voice actors filed a brief backing the New York Times in its lawsuit against OpenAI and Microsoft, with statements of support from Anne Hathaway, Emma Thompson, Ron Howard and Madonna. OpenAI says licensing the world’s creative work to train its machines would cost too much. The stars reply that “the creative industries already have in place a proven infrastructure and workflows to clear rights for dispersed works at scale and cost-effectively”, in the words of the brief.

Blanchett put it in one sentence: “If someone wants to use my face, my voice or my image, they ask me first.”

Read closely, it is the language of a settlement. After three years of strikes, picket lines and lawsuits, Hollywood’s talent has concluded that it cannot defeat the machine, so it wants to be paid by it. The plaintiff asks for consent and a fee, and AI becomes a new source of income. Writers and musicians are learning to license their words and songs. The actors will be next, with their faces.

The Price of a Digital Double

The voice actors have already shown how it will work. On Thursday SAG-AFTRA members ratified new television animation contracts with 94.95 per cent of the vote, and the union’s president, Sean Astin, hailed “a dignified deal with essential AI protections”. The protections are real. From productions starting in July 2027, a studio will need an actor’s consent to use a digital replica to dub his performance into another language. A replica may not cross a picket line.

The fine print tells the bigger story. When a studio builds an actor’s voice from his recordings without him, the deal sets the minimum session fee as the minimum payment, plus residuals. The union has put a price on the double: one day’s pay at the microphone. Synthetic performers may still take human roles when they bring “significant additional value”, a phrase studio lawyers will savour.

SAG-AFTRA can slow the machine and set its rates. Stopping it is beyond any union, and the stars are already queueing. Michael Caine put his voice on ElevenLabs’ marketplace almost a year ago, promising to open “doors for new storytellers everywhere”. Disney invested $1 billion in OpenAI and licensed more than 200 of its characters to Sora, in a deal that leaves out “talent likenesses or voices”, for now. Tilly Norwood, the computer-generated “actress”, has been cast as the lead in a feature film, Misaligned. The machine can already render a face the eye cannot tell from the real one.

What the Machine Cannot Remember

Hollywood has already tried the trick on its greatest actor. In The Irishman, digital effects stripped decades from Robert De Niro’s face so that, at 76, he could play Frank Sheeran as a young man. Critics complained that he moved like an old man wearing a young face. De Niro told UKScreen that a movement coach on set tapped the actors on the back whenever they slouched, to remind them they were supposed to be 30 or 40, and he conceded that small things slip through however hard you try. The computer rewound the face. The years lived on in the body, and the camera saw them.

Husam Sam Asi talking with Robert De Niro during promotion for The Irishman.
Husam Sam Asi with Robert De Niro during promotion for The Irishman. Photo: Armando Gallo.

What the eye can tell, without knowing why, is a performance. An actor’s reaction to a moment rises from the subconscious, and it draws on the life the actor has lived. AI has a prodigious memory for knowledge. It has no emotional memory. Konstantin Stanislavski built modern acting on that faculty a century ago, teaching actors to summon feelings from their own past to bring a scene to life. Lee Strasberg made it the heart of the Method.

Husam Asi with Daniel Day-Lewis during promotion for Lincoln.
Husam Sam Asi with Daniel Day-Lewis during promotion for Lincoln.

The great actors describe the same process again and again. The real work of becoming Lincoln, Daniel Day-Lewis told me, was “finding those corresponding elements in your life with the life of the character”. To find Sheeran, De Niro looked for the similarities between the hitman and himself: “Make it your own, personalise it,” he explained. Al Pacino put it in a line: “There’s always a part of you in a part – but just a part.” Gary Oldman read Churchill’s books and watched the footage, then went further: “Reading a book doesn’t give you that. You have to work and find it through feeling.”

The machine has read more books than any actor alive. It has nothing to feel with.

Blanchett Against Her Avatar

The best case against Blanchett’s digital double comes from Blanchett herself. In 2013, talking about Blue Jasmine, she told me that an actor can craft a whole scene at home, every line and gesture, and then, when the other actor does something different, “you respond differently”. The secret, she said, is that “you have to be alive to what happens in the moment”.

An avatar has no moment to be alive in. It predicts the likeliest reaction from millions of performances it has swallowed. A great actor gives the reaction nobody predicted, because it comes from a grief, a joy or a humiliation only she has lived. SAG-AFTRA said as much when Tilly Norwood appeared: “It has no life experience to draw from, no emotion,” the union declared.

Selling the Shadow

That is the trap in the settlement. An actor who licenses an avatar sells a shadow: the face, the voice and the tics. Every licence also feeds the machine more of the performances it learns from, and teaches it to fake the rest. Studios drowning in debt will gladly take a shadow at the price of a session fee, and a talent that took a lifetime of living to build will be traded for residuals.

“Consent isn’t a barrier to creativity, it is the foundation of it,” Blanchett says. Consent protects the face. The talent behind it survives only if actors refuse to sign it away.

More on UKScreen

Most read on UKScreen

  1. Musk and China Handed Cinema to AI. Audiences Felt Nothing
  2. Netanyahu Couldn’t Stop NAZA. Now Hollywood Wants a Piece of It
  3. Washington hosted them, Hollywood Lionised them. Now They’re Terrorists
  4. Sacha Baron Cohen, Boycotter, Becomes the Boycotted as Ali G Returns
  5. Elizabeth Holmes Beat the Rich at Their Own Game. Now Cinemas Cheer Her
The UKScreen Letter

Get the best of UKScreen by email

New stories, reviews and interviews from the archive, straight to your inbox.