AI Learns to Sing? Publishers Sue Anthropic Over Lyrics

Three major music publishers have asked a California judge to rule that U.S. copyright law does not shield artificial intelligence startup Anthropic from liability. The publishers, including Universal Music Group, Concord, and ABKCO, allege that Anthropic copied song lyrics to train its AI-powered chatbot, Claude. The publishers filed a request on March 24, 2026, in federal court, seeking a crucial determination on whether 'fair use' applies to the mass copying of copyrighted works for AI model training.

The publishers contend that Claude's AI-generated lyrics are not fair use because they are derivative works that directly compete with and diminish the market for the original lyrics. Furthermore, they claim Claude unlawfully reproduces their lyrics on demand without permission. This lawsuit is one of many legal disputes between copyright holders and tech giants over the use of creative content in AI systems. Anthropic, backed by Amazon and Google, previously settled a $1.5 billion class-action lawsuit with a group of authors.

The music publishers initially sued Anthropic in 2023, asserting that their copyrights were infringed concerning lyrics from at least 500 songs by artists like Beyoncé and the Rolling Stones. The central issue in these cases is whether AI systems engage in fair use by employing copyrighted material to create new, transformative content. The publishers specifically urged U.S. District Judge Eumi Lee to rule on copyright infringement before a trial, rejecting Anthropic's fair use defense. This comes as differing judicial opinions on fair use in AI training have emerged in Northern California courts.