Anthropic Reaches Agreement with Music Publishers on Copyright Issues

The copyright dispute between music publishers and AI company Anthropic has finally reached a stage of resolution. Under the latest agreement reached, Anthropic will take steps to ensure that its AI chatbot, Claude, does not infringe on copyrights when providing lyrics. Under the latest agreement reached, Anthropic will take steps to ensure that its AI chatbot, Claude, does not infringe on copyrights when providing lyrics. The move provides a short-term solution to the ongoing litigation and shows some sort of compromise between the two sides on this complex issue.

Back in October 2023, Universal Music Group, ABKCO, Concord Music Group, and a number of other music publishers jointly filed a lawsuit against Anthropic, alleging that it utilized the lyrics of at least 500 copyrighted songs to train its artificial intelligence system. The lawsuit materials show that when Claude was asked to provide lyrics to songs such as Beyoncé's “Halo,” Mark Ronson's “Uptown Funk,” and Maroon 5's “Moves like Jagger,” the chatbot responded with content that “almost exactly replicated or contained key lyrical portions of those songs “.

While music publishers acknowledge that platforms such as Genius have made lyrics publicly available online, they emphasize that these sites pay licensing fees for the use of copyrighted works. The lawsuit states that Anthropic “intentionally removed or altered” the “copyright management information” of the affected songs when it collected data from the sites to train its artificial intelligence models.

Under Thursday's agreement, Anthropic promised to continue to enforce its existing copyright protection measures to prevent its AI models from infringing on copyrighted content, and to apply those measures to future AI systems it develops. In addition, the music publisher and Anthropic will work together “in good faith” to resolve any protections that may be deemed ineffective, while the court reserves the right to resolve any disputes.

In a statement, Anthropic said, “Claude's designs were not intended to be used for copyright infringement, and we have implemented processes to prevent such infringement. Our decision to enter into this agreement is consistent with these priorities of ours. We remain convinced that the fair use of potentially copyrightable material in the training of generative AI models is lawful under current copyright law.”

Meanwhile, the music publishers behind the plaintiffs have asked the court to grant a preliminary injunction prohibiting Anthropic from using its protected lyrics for model training in the future, which the court is expected to rule on in the next few months.