January 4 news.Anthropic The company works with majormusic publisherreached an agreement to ban its AI assistants by Claude generating copyrighted lyrics, temporarily resolving the dispute between theHowever, the core issue of "using copyrighted lyrics to train AI models" remains unresolved.
Anthropic reached an agreement with music giants such as Universal Music Group, Concord Music Group and ABKCO that prohibits Claude from copying or creating lyrics based on copyrighted material.
Note: The agreement stems from the publisher's 2023 lawsuit against Anthropic, in which Claude was accused of copying the lyrics to at least 500 songs involving a wide range of artists, including Katy Perry, The Rolling Stones and Beyoncé.
Anthropic says that Claude was not designed to infringe on copyrights and that the company has already taken appropriate protective measures, which are further strengthened by this agreement, under which Claude cannot reproduce copyrighted lyrics or create new lyrics based on protected material.
Anthropic has also instituted a new notification mechanism: if a publisher becomes aware of a Claude violation, it can notify Anthropic in writing, and the company promises to deal with it promptly.
While the agreement resolves the dispute over Claude's output of lyrics, the central question of whether AI companies can legally use copyrighted lyrics to train their models remains unanswered. anthropic, like other AI labs, maintains that this is "fair use," but the courts have yet to rule definitively on this issue. The courts have yet to make a definitive ruling on the issue.
