Judges, Guardrails and Ongoing Litigation
In the earlier case, U.S. District Judge Eumi K. Lee ruled last October that amended claims against Anthropic were sufficient to support allegations of contributory and vicarious infringement. She rejected the company’s argument that the publishers failed to show Anthropic knew — or was willfully blind to — how users might infringe copyrights using Claude.
Last month, Anthropic sought a court order compelling publishers to turn over the prompts and outputs used by investigators to test whether Claude reproduced song lyrics. The publishers criticized the request as overly broad and a renewed attempt to pierce work-product protections.
Parallel Fights in the AI Copyright Wars
The music case unfolds alongside other high-profile legal battles facing Anthropic. In a separate lawsuit brought by bestselling authors accusing the company of using their works to train AI systems, Anthropic agreed to pay $1.5 billion to resolve the claims.
The music publishers argue that, rather than changing course, Anthropic has intensified its conduct since the first lawsuit was filed — scraping data online, copying physical books and exploiting datasets that allegedly contain unauthorized copies of copyrighted music.
They contend Anthropic trains its models to memorize works, leading to outputs that replicate copyrighted material in multiple ways. Beyond that, they say, the AI generates vast quantities of new, AI-written lyrics derived from copyrighted songs, competing directly with legitimate works and harming songwriters.
