Anthropic sued by music industry giants over allegations of copyright misuse in training its AI model

Photo by Elizeu Dias on Unsplash

Music publishing giants Sony ‌Music and Warner are claiming Artificial Intelligence (AI) pioneer Anthropic ‘pirated’ their copyright songs to train its Claude family of advanced Large Language Models (LLMs). 

The law suit, filed in California federal court, follows Anthropic’s agreement to pay $1.5bn to settle a similar court action over claims that it illegally accessed copyright books to train Claude. 

In the book case, a class action brought by authors and publishers, the judge ruled in June 2025 that, while it was not illegal to use copyright material to train AI, Anthropic should face trial over illegally obtaining books from pirate websites.  

As the BBC reported last September, the judge in the case, Judge Alsup, said in his June decision that Anthropic “holds more than seven million pirated books in a central library” and that the company “faced up to $150,000 in damages per copyrighted work.” 

The settlement was for considerably less than that, with authors reportedly receiving around $3,000 per book used.  

In the latest complaint, filed on August 28, 2026, Sony and Warner allege: 

“Defendants Anthropic and its founders Dario Amodei and Benjamin Mann have conducted a brazen campaign of illegally torrenting, scraping, and downloading copyrighted works on a massive scale in order to develop, operate, and reap enormous profits from Anthropic’s “Claude” series of artificial intelligence (“AI”) models.” 

The complaint claims that Anthropic illegally accessed “tens of thousands of copyrighted musical compositions and song lyrics,” including works by The Beatles, Taylor Swift, and Bon Jovi. Significantly, the case appears to rest on claims that it was written material and books of songs that were ‘pirated,’ not recordings or videos: 

“These books contain properly licensed copies of Music Publishers’ sheet music and song lyrics and CMI [Copyright Management Information – details of a work’s title, author, copyright holder, terms of use etc] attributed to each work. These books also typically state on their face that unauthorized copying of the works contained therein constitutes copyright infringement—a warning Defendants blatantly ignored.” 

The music labels are seeking damages of up to $150,000 for each infringed copyright and a court order barring Anthropic from using their works. The $150,000 figure frequently appears in law suits over copyright – it is the maximum statutory damages per infringement allowed under US law. However, damages can be much higher in copyright cases where claimants ask for a figure based on the actual damages they suffered and the profits made through infringement, instead of the statutory amount. 

The fact that the latest filings make extensive references to evidence of piracy submitted in the previous book copyright class action strongly suggests that Sony and Warner’s lawyers see this as the main thrust of their argument, given Judge Alsup’s June ruling that simply using copyright material to train an LLM was not in itself illegal. 

A Reuters story on the latest music court case reports:  

"Anthropic clearly considers [the $1.5bn book copyright settlement] to be just the cost of doing business given that its entire business model continues to be built on copyright theft," Sony and Warner said in their 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." 

According to Reuters, an Anthropic spokesperson responded: “This is the third lawsuit from the same lawyers, recycling allegations from cases already before the courts"… The spokesperson also said that the company will defend itself "robustly" and that ⁠AI training makes fair use of copyrighted material, citing [Judge Alsup] in the authors' case against it.” 

Sony and Warner join an extensive list of companies representing copyright holder who believe their works have been used to train AI models who are seeking redress through the courts in the US and elsewhere.