Artificial intelligence company Anthropic has been sued by two of the world’s major music publishers, which accuse the company of using tens of thousands of copyrighted songs without permission to train its Claude chatbot. Sony Music Publishing and Warner Chappell Music filed the lawsuit in a US federal court in California, escalating a growing battle between AI companies and copyright holders over the use of creative works to develop generative AI systems.
The publishers allege that Anthropic illegally obtained and copied large quantities of copyrighted lyrics and musical compositions and used them as training material for its artificial intelligence models. According to the complaint, the company allegedly relied on pirated sources, including torrent downloads, as well as scraping material from online platforms containing song lyrics. The lawsuit claims the works included music associated with some of the world’s best-known artists, including The Beatles, Taylor Swift and Michael Jackson.
The complaint also identifies numerous well-known songs that the publishers say were among the copyrighted works used by Anthropic. These include Mariah Carey’s “All I Want for Christmas Is You”, Bon Jovi’s “Livin’ on a Prayer”, “Eye of the Tiger”, “Ain’t No Mountain High Enough”, “September” and Taylor Swift’s “Paper Rings”. The publishers argue that Anthropic’s alleged conduct was not limited to a small number of works but involved a vast collection of music. The lawsuit reportedly covers at least 20,000 copyrighted works, potentially exposing the company to damages running into billions of dollars.
Sony and Warner allege that Claude has been trained in a way that allows it to reproduce copyrighted lyrics when users specifically prompt the system. They further argue that the AI can use material from existing songs to generate large quantities of new lyrics, creating products that could compete with original music and reduce the commercial value of works created by songwriters and composers. The publishers contend that such use goes beyond legitimate technological development and amounts to infringement of their intellectual property rights.
The lawsuit is also significant because of allegations concerning the way copyright information was handled. The music companies claim Anthropic removed identifying information associated with some copyrighted works. Such information can identify creators and rights holders and is important for attribution and licensing. The plaintiffs are seeking statutory damages of up to $150,000 for each infringed work, along with additional damages relating to the alleged removal of copyright-management information.
The case adds to a series of copyright disputes facing Anthropic and other leading AI developers. Anthropic has already faced legal action from authors and music companies over the use of copyrighted material to train its AI models. In July, a US federal judge approved a $1.5 billion settlement in a separate case brought by authors who accused Anthropic of using their books to train Claude. That settlement was described as the largest known payout in a US copyright case.
Music companies have also previously challenged Anthropic’s use of song lyrics. Universal Music Group and other music publishers filed a separate lawsuit against the company in 2023, alleging that copyrighted lyrics were being used in AI training. That case remains part of the broader legal debate over whether AI developers can use copyrighted material to train their systems without obtaining licences or paying rights holders.
The latest lawsuit could therefore become another important test of copyright law in the rapidly developing AI industry. At the heart of the dispute is the question of whether companies developing generative AI can copy and analyse enormous quantities of copyrighted creative works to train their models, and under what circumstances such use may qualify as fair use.
Anthropic has faced increasing legal and regulatory pressure as its AI business has expanded. The company has denied the allegations made against it and is expected to contest the claims in court. Sony Music and Warner Chappell, meanwhile, are seeking an injunction that would prevent Anthropic from continuing to use their copyrighted works in the manner alleged in the lawsuit.
The case highlights the widening conflict between the rapid expansion of generative AI and the interests of artists, authors, publishers and other copyright holders. As AI companies compete to build increasingly capable systems, they require enormous quantities of data for training. Rights holders, however, are increasingly demanding that their works cannot be used commercially without permission, attribution or compensation. The outcome of the Anthropic case could therefore have implications well beyond the music industry, potentially influencing how AI companies obtain and use copyrighted material in the future.
