In a significant legal development affecting the artificial intelligence industry, US District Judge Eumi K. Lee has largely allowed a copyright infringement lawsuit against AI company Anthropic to proceed while rejecting music publishers‘ attempts to expand the case with new piracy allegations.
The judge denied Anthropic’s motions to dismiss, reinstating full copyright claims that had previously been partially dismissed, finding the publishers’ revised arguments “plausible” and “sufficient” to move toward trial.
The lawsuit, filed in 2023 by Universal Music Publishing Group, Concord Music, and ABKCO, alleges that Anthropic’s Claude chatbot was trained on copyrighted music lyrics without permission.
The case involves more than 500 songs from artists including Beyoncé, the Rolling Stones, and the Beach Boys, with publishers claiming the AI system reproduces their protected lyrics when prompted by users.
Judge Lee specifically rejected the publishers’ attempt to amend their complaint with allegations that Anthropic had used pirated materials, such as content obtained via BitTorrent, to train its AI model.
Despite publishers’ efforts to add BitTorrent piracy claims, Judge Lee rejected these new allegations for lack of supporting evidence.
The court found these new piracy claims lacked adequate investigation and evidence, despite similar allegations surfacing in a separate lawsuit involving book authors that recently resulted in a $1.5 billion settlement by Anthropic.
The ruling addresses multiple aspects of copyright liability, including direct, contributory, and vicarious infringement.
Significantly, the judge highlighted that Anthropic’s own safety “guardrails” demonstrated the company’s awareness that users were generating copyrighted lyrics, potentially establishing the knowledge requirement for secondary liability claims.
Amazon-backed Anthropic, which recently secured $13 billion in investment, now faces continued litigation over whether AI companies can be held responsible for both their training data practices and user-generated outputs.
The publishers involved are represented by performing rights organizations that typically collect and distribute royalties for public performances of their members’ copyrighted works.
The music industry has been increasingly focused on protecting artists’ rights across streaming platforms as digital distribution becomes the dominant revenue channel for most musicians.
The copyright management information claims also remain active, with publishers alleging Anthropic removed digital CMI from protected works during their AI training process.
The decision, issued on October 8, 2025, represents another chapter in the growing number of legal challenges facing major AI developers and tech companies over copyright concerns.
The case represents an important precedent as courts navigate the intersection of copyright law and artificial intelligence, with Judge Lee’s decision maintaining the focus on established copyright claims rather than expanding into more speculative piracy allegations.
