Music Publishers Sue Anthropic Alleging 'Brazen' Copyright Theft in Claude Training

The music industry's biggest publishers have taken aim at Anthropic. In a lawsuit filed late Friday in the U.S. District Court for the Northern District of California, Sony Music Publishing, Warner Chappell, and a coalition of publishers allege that Anthropic — and co-founders Dario Amodei and Benjamin Mann — engaged in a "brazen campaign of illegally torrenting, scraping, and downloading copyrighted works" to train its Claude models.

The publishers accuse Anthropic of "blatant theft," claiming the company used thousands of copyrighted works, including those containing lyrics and sheet music, without authorization. The suit, first reported by Music Business Worldwide, builds on earlier cases but is notably broad in scope, explicitly calling out "flagrant piracy" through illegal torrenting.

This isn't Anthropic's first rodeo. The same legal team previously brought the Bartz v. Anthropic case, where a court ordered Anthropic to pay $1.5 billion after ruling that while using copyrighted works for training can be legal, acquiring them through piracy is not. That precedent looms large over the new suit.

Anthropic had not responded to requests for comment at press time. The outcome of this case could have significant implications for how AI labs source training data — and whether the industry's reliance on large-scale scraping is sustainable.

For developers and technologists, this case underscores the legal risks lurking in training data pipelines. Even if your model's output is transformative, the method of acquisition matters. As courts continue to weigh in, expect the cost of data to rise — and the era of unencumbered scraping to end.