Sony Music Publishing and Warner Chappell Music filed a sweeping copyright lawsuit against Anthropic on August 28, accusing the company of a brazen campaign of illegally torrenting, scraping, and downloading copyrighted songs to train its Claude models. What sets this suit apart from the dozens of similar AI copyright cases filed over the past three years is who else is named alongside the company: CEO Dario Amodei and co-founder Benjamin Mann, personally, as individual defendants.
What the complaint actually alleges
The 48-page complaint, filed late on August 28 in the U.S. District Court for the Northern District of California, accuses Anthropic of obtaining and using tens of thousands of copyrighted musical compositions without permission or payment, drawing on well-known songs including Eye of the Tiger, Marvin Gaye’s Ain’t No Mountain High Enough, Uptown Funk, Hallelujah, Mariah Carey’s All I Want for Christmas Is You, and Taylor Swift’s Paper Rings. The publishers describe the alleged conduct as “one of the largest and most blatant ongoing thefts of intellectual property in history,” and say Anthropic obtained the material through torrent networks and large-scale scraping, pulling from digital archives including Library Genesis and Pirate Library Mirror, the same sources at the center of Anthropic’s earlier book-piracy litigation.
The complaint cites internal Anthropic material that was unsealed during that earlier case, including a description from Mann himself calling Library Genesis “sketchy AF,” and a characterization from Anthropic’s own internal Archive Team describing the same source as a “blatant violation of copyright.” The publishers are seeking statutory damages of up to 150,000 dollars for each work found to have been willfully infringed, plus up to 25,000 dollars for each instance of alleged removal or alteration of copyright management information, a combination that could theoretically push total exposure into the billions of dollars depending on how many works a court ultimately finds were infringed. They’ve also requested a jury trial and remedies including destruction of infringing copies and a full accounting of what training data actually went into Claude.
Why naming the founders personally is the real story
Naming individual founders as defendants in a corporate copyright case is genuinely rare, and the complaint is explicit about why the publishers are doing it here: it alleges Amodei and Mann are personally liable for their specific roles in directing the illegal torrenting, framing them as contributory infringers who knew about and actively drove the large-scale use of pirated material, not merely executives overseeing a company that happened to do so. The practical effect of that legal strategy is significant. Claims against individual defendants can proceed independently of whatever ultimately happens to the company itself, whether Anthropic settles, restructures, or is acquired down the line. It also imposes a different, more personal category of legal risk on Anthropic’s leadership than ordinary corporate litigation typically does, and shifts the negotiating dynamic in ways a purely corporate defendant wouldn’t face.
The full weight of the music industry, now aligned
This lawsuit means the publishing arms of all three major music companies are now actively litigating against Anthropic. Universal Music Publishing Group, Concord Music Group, and ABKCO originally sued Anthropic in Nashville back in October 2023 over roughly 500 songs, a case later transferred to California, and BMG filed its own separate lawsuit alleging infringement of 493 compositions. The Sony and Warner Chappell suit is notably broader than either of those, alleging infringement across tens of thousands of compositions rather than a narrower, specifically enumerated list.
The case also follows directly on the heels of a legal precedent that likely emboldened the filing: a Munich court ruled in November 2025 that a model memorizing song lyrics constitutes reproduction under copyright law, and that the text and data mining exception that AI companies frequently lean on for training-data defenses does not cover that kind of memorization. Anthropic has already been through one massive reckoning on this exact underlying question. In September 2025, the company agreed to a 1.5 billion dollar settlement with authors and publishers, the largest copyright settlement in U.S. history at the time, stemming from separate allegations that it trained Claude on more than 7 million pirated books.
Anthropic’s response, and what happens next
An Anthropic spokesperson told reporters the company disagrees with the publishers’ claims and intends to defend itself robustly in court, without addressing the specific allegations against Amodei and Mann individually. Given the scale of Anthropic’s prior book-related settlement and the specificity of the internal documents already cited in this new complaint, the case looks likely to become a genuine test of how far personal liability for AI training decisions can actually extend to a company’s founders, not just its corporate entity, a question with implications well beyond Anthropic or the music industry specifically.
See Music Business Worldwide’s full report, which includes the complete complaint.
This case follows a similar pattern to our earlier coverage of how AI companies are handling large-scale legal exposure over training data sourcing.
