A coalition of music publishers, including UMG, filed a second lawsuit against Anthropic, seeking $3B over allegations of copyright infringement of 20k+ songs
Context & Ripple Effects
Music publishers’ dispute with Anthropic began with a 2023 lawsuit alleging lyrics were used to train AI models. The new action broadens the pressure from an initial challenge into a larger claimed infringement dispute involving more than 20,000 songs.
The escalation comes alongside reports that major music companies were exploring licensing talks with AI music startups. That contrast makes this case consequential: publishers are pursuing negotiated access in some parts of AI while litigating alleged unlicensed use in others.
First-order effects
- Anthropic faces a second copyright claim from a publisher coalition, with $3 billion sought over alleged infringement involving 20,000-plus songs.
- The publishers, including UMG, gain another vehicle to press Anthropic over the use of music works and to seek damages rather than rely solely on voluntary licensing discussions.
Second-order effects
- AI companies handling music or lyrics will face stronger incentives to document training-data provenance and negotiate rights where they cannot establish authorization.
- Music rightsholders’ bargaining leverage in AI licensing discussions may increase, as the claimed scale of alleged use is put before a court.
Third-order effects
- If courts continue to entertain large-scale training-data claims, music AI may move toward a more formal market for licensed training access, with rights clearance becoming a product and operating constraint.
- The split between licensing negotiations and infringement litigation points to a durable industry question: whether AI developers can secure scalable rights arrangements before products and datasets are deployed.
The trend: Music rightsholders are using both licensing negotiations and copyright litigation to shape the terms on which AI systems can use songs and lyrics.