Sony Music, Warner sue Anthropic, alleging a “brazen campaign” of intellectual property theft

Music publishers led by Sony Music Publishing and Warner Chappell have brought a new copyright case against Anthropic, accusing the artificial intelligence company of obtaining protected material through unlawful means for use in developing Claude. The complaint also names Anthropic co-founders Dario Amodei and Benjamin Mann. Filed in federal court in Northern California late Friday, the action puts the dispute over AI training data squarely on allegations of piracy as well as copyright infringement.

The publishers say Anthropic pursued copyrighted works on a large scale, describing conduct that included torrenting, web scraping and direct downloads. Their central allegation is that the company acquired and used thousands of protected works while training Claude, its AI model. The plaintiffs characterize the alleged conduct as an intentional appropriation of material that publishers own or administer, rather than a narrow disagreement over a small set of examples.

Music Business Worldwide first reported the filing. Its arrival adds to the growing legal pressure facing companies that build generative AI systems from very large collections of text and other material. The case is notable because the publishers’ account focuses not simply on what an AI system may produce after training, but on the route by which copyrighted works were allegedly collected before that training occurred.

That distinction could matter as the lawsuit develops. Copyright disputes involving AI often raise questions about whether using protected works to train a model requires permission, compensation, or both. Here, the publishers are also challenging the claimed acquisition process itself. If those assertions are tested in court, the litigation may require close examination of how datasets were assembled, where material came from, and what role particular methods of collection played in model development.

Sony Music Publishing and Warner Chappell are among the best-known businesses that represent songwriters and manage musical compositions. Their participation, alongside a number of other publishers, gives the suit broad industry backing. The complaint concerns publishers’ rights in copyrighted works; it does not, based on the available account, resolve separate questions about recordings, artists, or any specific output generated by Claude.

Anthropic rejected the claims in a statement provided by email. “We disagree with the publishers’ claims,” a company spokesperson said, adding that Anthropic plans to contest the case vigorously. The response sets up an early and fundamental conflict between the plaintiffs’ depiction of systematic misconduct and the company’s position that the allegations lack merit.

For the music business, the suit reflects a larger concern that the rapid expansion of AI tools could alter how valuable creative works are used without the participation of the people and companies that control them. Publishers have a commercial interest in protecting compositions and licensing their use. AI developers, meanwhile, have faced mounting scrutiny over the sources and legal status of the material used to build systems capable of generating and analyzing language.

The case is still at an early stage, and the allegations in the complaint have not been proven in court. Anthropic will have an opportunity to respond formally, and the parties could dispute both the factual account of how material was obtained and the legal consequences of any use in training. A court will also need to determine which claims can move forward and what evidence is relevant. No outcome is assured merely because a complaint has been filed, and the company’s defense may challenge the publishers’ narrative at several points.

What happens next will be watched beyond the immediate parties. A ruling that addresses alleged torrenting, scraping, or downloading could shape how AI companies assess data-collection practices, while a decision focused on training uses could influence future licensing talks with rights holders. The case may also sharpen attention on whether businesses developing AI have documented, legally defensible methods for gathering material. For now, the lawsuit underscores that the debate over generative AI and copyright is moving from broad policy arguments into detailed disputes over individual companies’ practices.

Source: TechCrunch