Sony Music Publishing, Warner Chappell Music, and 35 affiliated music publishers filed a copyright infringement lawsuit against artificial intelligence startup Anthropic, Chief Executive Officer Dario Amodei, and co-founder Benjamin Mann on Friday, August 28, 2026. According to the court filing in the U.S. District Court for the Northern District of California, the music publishers accuse the company of orchestrating a systematic campaign of unauthorized data ingestion to train its Claude artificial intelligence models.
The complaint alleges that Anthropic engaged in illegal torrenting, scraping, and downloading of copyrighted musical compositions on a massive scale. Per the legal filing, the startup utilized piracy sources including Library Genesis and Pirate Library Mirror, where it allegedly accessed 5 million and 2 million pirated books respectively, alongside licensed lyric websites such as Musixmatch and LyricFind. The publishers assert that the company pirated hundreds of song lyrics and sheet music works from notable artists including The Beatles, Taylor Swift, and Michael Jackson to train Claude in generating text responses.
The music publishers are seeking statutory damages of up to $150,000 for each willfully infringed copyright, alongside an additional $25,000 for each alleged instance of copyright management information removal. While the exact number of specific musical works involved has not been fully itemized in public complaint summaries beyond thousands of compositions, the plaintiffs emphasized the economic scale of the alleged infringement.
"Anthropic clearly considers that to be just the cost of doing business given that its entire business model continues to be built on copyright theft," Sony Music and Warner Chappell stated in their joint complaint. "And $1.5 billion is obviously not a large enough settlement to deter infringing conduct by a company that has parlayed such mass infringement into a staggering $2-trillion-dollar valuation."
Anthropic Defends Training Practices Under Fair Use
An Anthropic spokesperson responded to the litigation during a statement issued on Monday, September 1, 2026, confirming that the startup intends to defend itself against the allegations.
"Anthropic will defend itself robustly," the company spokesperson stated, arguing that artificial intelligence training constitutes fair use of copyrighted material. The representative characterized the new action as the third lawsuit brought by the same legal representation, stating that the complaint merely recycles allegations from ongoing cases already before the federal courts.
Legal experts note that the fair use doctrine serves as a primary defense for generative artificial intelligence developers, who argue that utilizing existing works to train foundational models is transformative and does not compete with or supersede the original commercial product.
Precedent and Broader Industry Stakes
The litigation builds upon an escalating series of intellectual property disputes involving the startup. Universal Music Group previously filed lawsuits against Anthropic in October 2023 and January 2026 over similar allegations regarding copyrighted song lyrics, with those proceedings remaining active in federal court. Furthermore, Anthropic reached a $1.5 billion class-action settlement with a group of authors in September 2025 to resolve separate claims concerning the unauthorized use of pirated books for model training.
The dispute shares legal parallels with other high-profile intellectual property actions in the technology sector, such as Getty Images versus Stability AI, where Getty accused the company of using over 12 million images without authorization to train its Stable Diffusion model.
The resolution of the Sony and Warner litigation carries significant implications for the generative artificial intelligence industry. Content creators across the music, publishing, and visual arts sectors are seeking binding judicial precedents that establish mandatory licensing requirements and compensation structures for copyrighted data utilized in commercial machine learning pipelines.
