Sony Music and Warner Chappell Sue Anthropic Over Alleged Copyright Piracy in Training Claude

Major music publishers, including Sony Music Publishing and Warner Chappell Music, have initiated legal proceedings against artificial intelligence startup Anthropic and its executive leadership, alleging the company systematically infringed on copyrighted musical works to train its AI systems. The complaint explicitly names co-founders Dario Amodei and Benjamin Mann alongside the organization.

Submitted late Friday in the U.S. District Court for the Northern District of California, the legal filing claims the San Francisco-based enterprise ran an unauthorized data extraction campaign involving scraping, torrenting, and downloading protected content. The plaintiffs accuse Anthropic of widespread copyright infringement by utilizing thousands of protected compositions to build its primary AI model, Claude, as first reported by TechCrunch AI based on initial reporting from Music Business Worldwide.

The lawsuit builds upon a growing wave of intellectual property disputes centered on the data collection practices of generative AI developers. The music publishers argue that Anthropic's methods breached federal copyright protections by accumulating proprietary material without securing permission or executing licensing agreements with rights holders.

A central component of the new lawsuit focuses on allegations that Anthropic employed peer-to-peer torrenting networks to collect massive digital archives. The plaintiffs assert in court filings that these downloads contained millions of digitized books that included protected song lyrics and formal sheet music.

The legal team representing the music publishers includes attorneys who previously managed high-profile copyright actions against the AI developer. Counsel involved in this action previously represented Concord Music Group and Universal Music Group in litigation initiated in January, while also directing the Bartz v. Anthropic class action suit brought on behalf of published authors.

In the Bartz case, the court established a critical legal precedent concerning artificial intelligence training sources. While the presiding judge ruled that using copyrighted text to train AI models is legally permissible under fair use principles, the court determined that acquiring that underlying data through pirated channels or illicit downloads is illegal, culminating in a $1.5 billion penalty against Anthropic.

Leveraging that prior ruling, the music publishers' new complaint targets the specific supply chains Anthropic used to gather its datasets rather than the abstract concept of machine learning training. The suit seeks financial damages and court injunctions to halt the unauthorized use of the disputed catalog items.

Anthropic did not immediately respond to requests for comment regarding the federal court filing. The company has historically defended its dataset creation as protected under fair use guidelines as it continues expanding its Claude software ecosystem.