Sony Music and Warner Chappell Seek Billions in Damages From Anthropic Over Alleged AI Training Piracy — BigGo Finance

Two of the largest music publishing groups in the world are taking aim at Anthropic with a sweeping copyright lawsuit that could expose the AI developer to billions of dollars in liability. Sony Music Publishing and Warner Chappell Music filed a joint complaint late Friday in the US District Court for the Northern District of California, accusing the company and its two co-founders of running what the plaintiffs describe as “one of the largest and most blatant ongoing thefts of intellectual property in history.”

The 48-page filing names CEO Dario Amodei and co-founder Benjamin Mann as individual defendants alongside the company, an unusual move that signals the publishers intend to hold leadership personally accountable. The suit alleges that Anthropic illegally torrented, scraped, and downloaded copyrighted material on a massive scale to train its Claude series of AI models, and that the company stripped copyright management information from works in the process.

Sony and Warner are asking for up to $150,000 for each infringed work and an additional $25,000 for every instance in which identifiable copyright data was removed. With the complaint alleging that “tens of thousands” of copyrighted musical compositions were harvested, the maximum potential award could reach several billion dollars if the court sides with the publishers.

The complaint paints a detailed picture of how the alleged infringement took place. It claims Mann used BitTorrent to download more than five million pirated books, while other employees obtained at least two million more from Pirate Library Mirror, a notorious shadow library. The publishers also allege that Anthropic scraped lyrics from sites like MusixMatch and LyricFind, which had paid to license content from music labels.

Specific songs cited in the filing as appearing in Anthropic’s training data include Marvin Gaye and Tammi Terrell’s “Ain’t No Mountain High Enough,” Bon Jovi’s “Livin’ On a Prayer,” Earth, Wind & Fire’s “September,” Leonard Cohen’s “Hallelujah,” and Taylor Swift’s “Paper Rings.”

A Growing Legal Storm

The lawsuit is the latest in a series of high-profile copyright actions against Anthropic, and it arrives on the heels of a landmark settlement that reshaped expectations for AI-related copyright litigation.

In September 2025, Anthropic agreed to pay $1.5 billion to resolve the Bartz v. Anthropic case, in which a group of authors accused the company of using pirated copies of their protected works to train its AI models. The settlement was the largest copyright deal in US history at the time. A judge in that case ruled that while using copyrighted works for AI training could be legal under certain circumstances, acquiring that content through piracy was not.

Earlier this year, Concord Music Group and Universal Music Group filed their own suit against Anthropic, claiming the company illegally downloaded more than 20,000 copyrighted songs for training purposes and seeking more than $3 billion in damages. BMG and Round Hill Music have also brought separate actions. The BMG case is narrower in scope, alleging infringement of 493 compositions, which highlights how much broader the Sony and Warner complaint is by comparison.

Some of the same lawyers behind the Sony and Warner filing also represent Concord and Universal in their case and led the Bartz litigation.

Anthropic’s Music Copyright Litigation at a Glance

Why Music Copyright Is Different

The music industry’s copyright structure is uniquely complex, and that complexity could work in the publishers’ favor. A single commercially released song can contain multiple copyrighted components — lyrics, sound recording, and composition — each potentially held by different rights holders, including artists, publishers, and labels. This fragmentation means a single song in a training dataset can give rise to multiple infringement claims.

Statutory damages are another factor. Under US copyright law, plaintiffs can seek statutory damages without having to prove actual financial losses, which lowers the evidentiary burden considerably. With thousands of works at issue and per-work damages capped at $150,000, the arithmetic can quickly escalate into the billions.

The publishers have requested a jury trial. Anthropic did not immediately respond to requests for comment.

Industry Implications

The outcome of this case could have far-reaching consequences for the broader AI industry. Music publishers have been among the most aggressive litigants in the copyright battles over generative AI, and a victory here would likely embolden further actions against other AI developers. The music industry’s history of aggressive copyright enforcement stretches back to the Napster and LimeWire eras, and the current wave of AI litigation represents a continuation of that posture in a new technological context.

For Anthropic, the stakes extend beyond potential damages. The company has positioned itself as a leader in AI safety and responsible development, and the allegation that it acquired training data through piracy directly challenges that reputation. The decision to name Amodei and Mann personally also raises the possibility of individual liability, a development that could change how AI executives approach data acquisition decisions.

The case is likely to proceed alongside the Concord and Universal suit, potentially creating parallel tracks of litigation that could pressure Anthropic toward settlement. With the $1.5 billion Bartz settlement as a recent precedent, the company has already demonstrated a willingness to resolve copyright claims through negotiation rather than prolonged courtroom battles.