Sony and Warner Sue Anthropic Over Alleged Mass Use of Copyrighted Songs to Train Claude

Sony Music Publishing and Warner Chappell Music sued Anthropic in federal court in Northern California over alleged unauthorized use of copyrighted musical compositions to train Claude. The complaint names Anthropic, CEO Dario Amodei and co-founder Benjamin Mann as defendants, and describes the alleged conduct as a large-scale campaign involving torrenting, scraping and downloading protected works.

The publishers are seeking damages tied to each infringed work, along with relief addressing alleged removal of copyright-management information. The case places Anthropic among a growing group of AI companies facing litigation from rights holders over the use of creative works in model development.

The complaint says Anthropic used tens of thousands of musical compositions, including lyrics and sheet music, without licensing permission. Works associated with artists such as The Beatles, Taylor Swift, Michael Jackson, Mariah Carey and Leonard Cohen are cited in coverage of the filing. The publishers argue that Claude's ability to reproduce or provide song lyrics reflects the underlying use of protected material.

The allegations focus on how Anthropic obtained and processed the works, rather than only on outputs generated for users. The-decoder described the complaint as claiming that Amodei directed, approved or controlled the conduct, while Variety highlighted the publishers' characterization of a “brazen campaign.” Those assertions remain allegations to be tested in court.

The financial exposure could be substantial because US copyright law allows statutory damages of up to $150,000 per infringed work when infringement is found to be willful. Fortune also reported a separate claim of $25,000 for each instance in which copyright-management information was removed. The complaint therefore treats the alleged conduct as a mass infringement case, with the number of works driving the scale of any award.

The publishers' language frames the dispute as a fight over the economics of generative AI. They contend that Anthropic built valuable commercial systems by using creative works without paying for them, while Anthropic's legal defenses are expected to address authorization, fair use, copying practices and the relationship between training inputs and model outputs. The case's outcome will depend on evidence and court rulings, not the complaint's rhetoric.

The Anthropic case extends a copyright conflict that already spans books, journalism, visual art and software. Music publishers have a distinct concern because lyrics and compositions are individually protected, commercially licensed and closely tied to royalty systems. The lawsuit arrives after other rights holders brought cases against technology companies over training data and generated content.

Anthropic also faced recent scrutiny over Project Panama, a plan described as involving destructive scanning of books for AI training. That episode gives the music litigation a broader context: publishers and authors are pressing companies to disclose data practices and negotiate permission before using large collections of protected works. Gizmodo characterized the filing as a return to the first wave of post-ChatGPT copyright lawsuits, while Avclub emphasized the delay between this case and earlier actions against AI firms.

The case raises immediate questions for AI companies that rely on large cultural archives. A ruling against Anthropic could increase demand for direct licensing agreements, stronger provenance systems and limits on training datasets containing lyrics, sheet music or other commercially controlled works. It could also raise costs for model development and alter how companies document permissions.

Music publishers have established licensing mechanisms, making the dispute especially consequential for negotiations between rights owners and AI developers. Companies building generative systems will face pressure to prevent memorization, preserve copyright information and respond to takedown demands. The allegations also create competitive stakes for other model providers, including OpenAI, Google and Microsoft, whose AI products depend on access to broad information and creative content.

The lawsuit will proceed through motions, discovery and arguments over Anthropic's data acquisition and training methods. The individual claims against Amodei and Mann add a personal dimension, but the central dispute concerns Anthropic's corporate conduct and the legal treatment of copyrighted music in AI training.

The case's practical effect will extend beyond any eventual damages award. A settlement could establish licensing terms or usage restrictions, while a court decision could clarify how existing copyright rules apply to training systems that ingest large collections and generate text in response to prompts. For songwriters, publishers and AI developers alike, the filing turns data provenance into a central business and legal issue.