Owners of Elvis Presley's intellectual property sued Anthropic on October 23, alleging the Claude developer copied lyrics from 274 Presley songs to train its models without permission and that the models can now generate unauthorized versions of the protected material1,2,3.
The complaint, filed by a subsidiary of Authentic Brands Group and other companies with stakes in Presley's music, seeks up to $41.1 million in statutory damages5. Authentic Brands Group controls Presley's intellectual property, including his name, image, and likeness rights.
The suit names songs including "Can't Help Falling in Love," "Don't Be Cruel," "Return to Sender," "All Shook Up," "A Little Less Conversation," and "(You're the) Devil in Disguise" among the 274 compositions at issue4. It accuses Anthropic of having "built its entire business model on infringing copyrights at an unfathomable scale".
The complaint describes instances in which users could prompt Claude to recreate and update Presley's songs. "These renders keep landing very close to the originals: same structure, same hooks reused with light rewording, and full sets of verses/choruses/bridges," the filing states. "That's edging past 'inspired by' into reproducing the copyrighted song, even when the words are swapped out".
The filing also alleges that Claude has continued to reproduce copyrighted material despite Anthropic's claimed safeguards to prevent its models from generating existing song lyrics or compositions.
An Anthropic spokesperson stated that the company disagrees with the claims and intends to defend the case fully.
The $41.1 million figure represents a requested statutory maximum, not a court award, and the allegations have not been tested in court.
ANALYSIS The suit puts two recurring questions in AI copyright disputes squarely at issue: the provenance of training data and the degree to which general-purpose models reproduce protected works in their outputs. The dual theory of liability, covering both the ingestion of copyrighted material during training and the generation of infringing outputs, tracks the structure of earlier music-industry complaints against AI developers.