Judge approves Anthropic’s $1.5 billion books settlement

A federal judge approved Anthropic’s $1.5 billion settlement with authors and publishers over books used to train its Claude chatbot. The deal works out to about $3,000 per eligible book. The Los Angeles Times reported that U.S. District Judge Araceli Martínez-Olguín granted final approval in San Francisco federal court. The agreement is non-reversionary, so unused money will not flow back to Anthropic, and it also allows additional $3,000-per-book payments for any class work above 500,000 added to the Works List by Anthropic. More than 482,000 books were covered by the ruling, and about 91% had already been claimed by authors or publishers. Plaintiffs’ lawyer Justin Nelson called it the largest known copyright recovery in history, while the judge said the settlement provided meaningful relief. The case began in 2024, when bestselling novelist Andrea Bartz and two other authors sued Anthropic. Last summer, Judge William Alsup ruled that training AI chatbots on copyrighted books was not illegal under fair use, but that Anthropic had wrongfully acquired millions of books through pirate websites and stored them. The settlement agreement is limited to past torrenting, scanning, retention and use of the covered works, including training, research, development and production of AI models. It does not release claims based on model outputs, future reproduction or future derivative works, and it does not create a licence to torrent, scan or train on copyrighted works. According to the settlement FAQ from UC Press, the deadlines to file a claim or opt out have passed. Payments are set to be sent directly by the settlement administrator in several rounds, with the first planned to begin in 2026 and additional rounds in 2026 and 2027.