A federal judge in San Francisco has approved a $1.5 billion settlement between Anthropic and a group of authors who sued the artificial intelligence company for using their books to train its Claude AI chatbot without permission.
U.S. District Judge Araceli Martinez-Olguin approved the deal on Monday, dismissing objections that the settlement amount was too low. The ruling comes as the end of what many see as the largest settlement ever in a U.S. copyright lawsuit.
It is one of dozens of lawsuits filed by authors, publishers and media companies against AI developers over their use of copyrighted material to train large language models. It’s also the first major AI copyright case in the U.S. to be settled.
The deal had previously obtained preliminary approval from then-U.S. District Judge William Alsup in September 2025 before his retirement.
The settlement follows a 2025 court ruling that training AI models with books was fair use under U.S. copyright law, Anthropic Deputy General Counsel Aparna Sridhar said.
“We came to this settlement in 2025, following the court’s landmark decision that training AI on books is fair use under copyright law—which remains the law today,” Sridhar said in a statement.
She said that more than 91 per cent of eligible authors and publishers have already claimed their share of the settlement money.
Lead counsel for the authors, Justin Nelson, called the settlement a landmark. “It is the largest known copyright recovery in history. “We look forward to making distributions to the Class as soon as practicable,” Nelson said.
The lawsuit, filed in 2024, said Anthropic used pirated copies of books to train its AI system, Claude, without authorisation. It is supported by Amazon and Alphabet.
In a landmark ruling in June 2025, Judge Alsup ruled that Anthropic’s training of AI on copyrighted books was fair use. But he also found that the company had violated copyright law by keeping a digital library of more than seven million pirated books, not necessarily only for the purpose of training A.I.
Another trial was set for December 2025 to determine damages for that alleged infringement. Potential liability was estimated at hundreds of billions of dollars before the parties reached the settlement.
