Sony Music Publishing, Warner Chappell Music and other publishers have filed a federal lawsuit accusing Anthropic of unlawfully obtaining and using copyrighted musical compositions in the development of its Claude artificial-intelligence models. The complaint also names Anthropic chief executive Dario Amodei and co-founder Benjamin Mann as defendants.

The case was filed in the U.S. District Court for the Northern District of California. The publishers allege that Anthropic used torrenting, scraping and downloading to acquire protected works, then copied material for model training and in generated outputs. Those assertions are allegations in a newly filed civil case, not findings by the court.

Anthropic disputes the publishers' account. A company spokesperson told TechCrunch that Anthropic disagrees with the claims and intends to defend itself in court. The litigation will therefore turn on evidence about how the material was obtained and used, as well as the legal treatment of training and output practices under copyright law.

Music Business Worldwide reported that the complaint identifies tens of thousands of compositions the plaintiffs say were infringed. It seeks a jury trial and statutory damages, including amounts that can reach $150,000 for each work found to have been willfully infringed and separate damages for alleged removal of copyright-management information. Those maximums do not represent an award, and the eventual exposure would depend on what claims survive and what a court or jury concludes.

The publishers cite well-known songs among the works they say were affected. Their theory is broader than a dispute over a small set of chatbot responses: they contend that Anthropic's acquisition process and repeated copying created infringement at multiple stages. Anthropic has not accepted that characterization.

The lawsuit joins a growing set of copyright cases involving the Claude developer. Other music publishers have brought separate claims, while authors previously challenged Anthropic's use of books. In the authors' litigation, a court distinguished between using works for training and acquiring copies through piracy, a division that is likely to remain important as new cases test different facts.

Sony and Warner's publishing arms are now participating directly in litigation over the materials used by generative-AI companies. The case also names individual founders, increasing the breadth of the defendants compared with a suit directed only at the company.

For the wider AI sector, the filing reinforces a central unresolved issue: a model developer may face separate legal questions about the source of its training corpus, the act of training, and what its products reproduce for users. Licensing agreements may address some disputes, but this complaint asks a federal court to assess conduct the publishers describe as unauthorized from the outset.

The next procedural steps will include Anthropic's formal response and decisions about which allegations can move forward. Until then, the complaint presents the publishers' version of events, while Anthropic's public position is that the claims are wrong and will be contested.