Artificial intelligence startup Anthropic has reached a settlement in a high-profile copyright class action lawsuit brought by a group of U.S. authors, becoming the first major AI company to resolve such a case.

The lawsuit, filed in San Francisco federal court, accused Anthropic of using millions of pirated books to train its AI assistant, Claude, without permission or compensation. Judge William Alsup previously ruled that while some AI training could fall under fair use, Anthropic may have crossed a legal line by saving as many as seven million books in a centralized library. That practice, the court said, could constitute copyright infringement.

Had the case proceeded, Anthropic faced potential liability in the billions of dollars, with U.S. copyright law allowing damages of up to $150,000 per work for willful infringement. A trial had been scheduled for December.

The company disclosed the settlement in a court filing on Tuesday, though the terms were not made public. Anthropic declined to comment, but the plaintiffs’ attorney, Justin Nelson, called the agreement “historic” and said details would be announced in the coming weeks. Judge Alsup has ordered both parties to submit requests for preliminary approval of the settlement by September 5.

The class action was led by writers Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who accused the Amazon- and Alphabet-backed startup of “wholesale piracy” to fuel its AI ambitions.

Legal scholars say the deal could shape the landscape for other ongoing copyright battles involving AI developers, including OpenAI, Microsoft, and Meta Platforms, which also face lawsuits from authors, publishers, and media organizations.

“The devil is in the details of the settlement and future litigation about the terms of the settlement,” said Shubha Ghosh, a professor at Syracuse University College of Law.

The outcome marks a significant moment in the fast-moving debate over whether AI firms can rely on copyrighted works to train their systems under the doctrine of fair use—or whether they must strike deals with rights holders.

For now, Anthropic’s settlement signals a potential shift toward negotiation rather than courtroom battles, even as the broader AI industry continues to grapple with escalating legal scrutiny.