Sony and Warner sue Anthropic over copyrighted song theft used for AI training

Sony Music Publishing and Warner Chappell Music have sued Anthropic, accusing the $2 trillion AI company of pirating copyrighted songs at massive scale to train Claude, its family of AI models.

The lawsuit, filed Friday in California federal court, alleges Anthropic obtained lyrics and sheet music through torrent downloads and other pirate sources rather than licensing the material. The works cited include songs from Taylor Swift, The Beatles, Michael Jackson, and hundreds of other artists, according to Reuters.

The publishers are seeking damages of up to $150,000 for each infringed copyright, a figure that could push the case into the billions of dollars depending on how many works are found to have been infringed. They are seeking a court order barring Anthropic from using their copyrighted material.

“The music publishing branches of Sony Music and Warner Music have sued Anthropic in California federal court for allegedly misusing their copyrighted ​song compositions to train its Claude AI models,” Reuters reported.

At the center of the case is a question hanging over the AI boom: Where did companies get the enormous amounts of data used to train their models, and did they have the right to use it?

Sony and Warner Accuse Anthropic of Pirating Music to Train Claude in Billion-Dollar Copyright Fight

Sony and Warner make one of the industry’s most aggressive accusations yet. The complaint alleges Anthropic engaged in large-scale torrenting and scraping involving tens of thousands of copyrighted works. It claims Claude can reproduce copyrighted lyrics “verbatim” in response to prompts and that Anthropic used those works to train Claude to create new AI-generated lyrics that compete with human-created songs.

“Anthropic clearly considers that to be just the cost of doing business given that its entire business model continues to be built on copyright theft,” Sony and Warner said in their complaint. “And $1.5 billion is obviously not a large enough settlement to deter infringing conduct by a company that has parlayed such mass infringement into a staggering $2-trillion-dollar valuation.”

Anthropic CEO Dario Amodei and co-founder Benjamin Mann are named as defendants in the case.

AI’s training data fight gets bigger

The lawsuit lands as courts wrestle with a growing stack of copyright cases involving AI companies and the material used to build their models.

Universal Music Group sued Anthropic in 2023 over the alleged use of copyrighted song lyrics in Claude’s training. That case remains ongoing. Authors, publishers, news organizations, and other copyright holders have brought similar claims against AI companies, making training data one of the industry’s biggest unresolved legal issues.

Anthropic has already paid heavily in another copyright dispute. The company agreed to a $1.5 billion settlement with authors in a class-action case involving books used for AI training. The Sony and Warner complaint points directly to that settlement in arguing that financial penalties have failed to deter the conduct they allege.

The new case goes beyond claims that copyrighted material was simply absorbed into a training dataset. Sony and Warner argue that Claude can reproduce protected lyrics and generate competing lyrics based on knowledge acquired from copyrighted works, potentially bringing both the acquisition and output sides of generative AI into the fight.

Anthropic disputes the allegations and is expected to defend its practices.

The stakes stretch far beyond Claude. A major victory for music publishers could increase pressure on AI developers to license copyrighted training material and provide clearer records of where their datasets came from. A victory for Anthropic could give AI companies more room to argue that training models on copyrighted works is permissible under existing law.

For an AI industry built on enormous datasets, the answer could reshape the economics of building the next generation of models.

Anthropic CEO