The brief is 31 pages, filed in the Southern District of New York on Friday and obtained by TheWrap. Its signatories are Cate Blanchett's Human Consent Foundation, the Creative Artists Agency, the Music Artists Coalition, the National Association of Voice Actors and the Creators Coalition on AI. Its appendix lists supporters including Anne Hathaway, Bette Midler, Emma Thompson, Helen Mirren, Dan Levy, Ron Howard, Madonna and Olivia Wilde.

It lands in The New York Times' copyright suit against OpenAI and Microsoft, and it attacks one specific plank of the defence: the claim that licensing creative work for AI training is impractical.

"Defendants have argued that licensing creative works for AI training is cost-prohibitive," the brief states. "Defendants also claim there is no settled means of valuing training data. The facts demonstrate otherwise. The creative industries already have in place a proven infrastructure and workflows to clear rights for dispersed works at scale and cost-effectively."

The mechanism it points to is the century-old apparatus of collective licensing: clearinghouses that administer whole repertoires through blanket licences and metadata-driven systems rather than negotiating every work individually, with CISAC and BMI named as examples. The argument is aimed squarely at the "dispersed owner" problem — the idea that training data has so many rightsholders that permissions cannot practically be obtained. If a working market already exists for exactly that shape of rights, the transaction-cost reasoning behind a fair-use defence gets considerably weaker.

Blanchett's own statement leans on the same point from the other direction, experience rather than economics: "I've been working in film and theater for over 30 years, and in that time, the principle has never changed. If someone wants to use my face, my voice or my image, they ask me first. There is a contract. There is a conversation. This isn't a new idea or a complicated legal framework, it is simply how we treat each other in this industry. Consent isn't a barrier to creativity, it is the foundation of it."

The filing came one day after SAG-AFTRA, the Writers Guild of America and the American Federation of Musicians submitted their own brief, urging the court "to uphold the basic principles of copyright law and reject a categorical fair use exemption for AI training in accordance with court precedent". Together they form a coordinated week for the creative industries, and they join a widening stack of friend-of-the-court filings — including one from the Wikimedia Foundation arguing that both sides of the case have the training-versus-output distinction wrong.

What the brief does not do is decide anything. An amicus brief is advocacy, not evidence; the court is under no obligation to adopt its framing, and OpenAI and Microsoft did not respond to requests for comment on it. The companies' broader position — that training on lawfully accessed material is fair use, and that the Times seeks to license something the law does not require — remains the thing the case will actually turn on.

But the brief does change the shape of the argument in public. "Licensing at scale is impossible" is a claim that can be tested against an industry that already licenses at scale; Hollywood has now put the receipts in front of the judge. The case was filed in December 2023 and has been grinding through discovery and motion practice since, so the practical effect of any single filing is limited. The political effect of a coalition that includes CAA, a voice actors' union and a list of A-listers saying the same sentence is not.

Sources