Sony Music Publishing and Warner Chappell Music have initiated a multibillion-dollar lawsuit against the artificial intelligence company Anthropic, alleging widespread unauthorized use of copyrighted songs to train its Claude AI models. Filed in a federal court in California on August 28, the lawsuit accuses Anthropic, along with CEO Dario Amodei and co-founder Benjamin Mann, of systematically downloading and using tens of thousands of copyrighted works without permission.
The complaint states that Anthropic engaged in “torrenting, scraping and downloading” lyrics and sheet music from millions of copyrighted works, including high-profile compositions such as Mariah Carey’s “All I Want for Christmas Is You,” “Ain’t No Mountain High Enough,” originally performed by Marvin Gaye and Tammi Terrell, Survivor’s “Eye of the Tiger,” and Bon Jovi’s “Livin’ on a Prayer.” According to the suit, at least seven million copies of books and music-related content were sourced from pirate websites, licensed lyric sites such as Musixmatch and LyricFind, and large datasets like Common Crawl.
The publishers contend that Anthropic stripped identifying information from the songs to prevent attribution, which they say denied music creators recognition and compensation. The companies describe the case as one of the most extensive and blatant intellectual property thefts in history, seeking damages that could amount to billions of dollars. The claim includes statutory damages of up to $150,000 per infringed work, with additional penalties for the removal or alteration of copyright information.
This legal action follows a previous settlement last year when Anthropic resolved a class-action lawsuit brought by U.S. authors, paying $1.5 billion over allegations of unauthorized use of copyrighted books to train its AI systems. Responding to the current allegations, Anthropic has denied any wrongdoing and stated its intention to defend itself vigorously in court, asserting that it has “nothing to hide” concerning its practices.
The case highlights ongoing legal and ethical challenges in the tech industry surrounding the use of copyrighted material to develop artificial intelligence models. As AI capabilities continue to expand, disputes over data sourcing and intellectual property rights have become increasingly prominent, involving authors, publishers, music labels, and news organizations worldwide. The coming litigation could set important precedents regarding the boundaries of AI training data and copyright law.