A federal judge in San Francisco has approved a $1.5B copyright settlement that forces AI company Anthropic to pay authors and publishers for using pirated books to train its Claude chatbot. The decision, issued by District Judge Araceli Martínez‑Olguín, provides “meaningful relief” to those whose works were used without permission.

Under the agreement, roughly 91% of the more than 482,000 books covered by the ruling have been claimed by authors or publishers, each receiving about $3,000 per book. Plaintiff attorney Justin Nelson called the settlement “the largest known copyright recovery in history,” adding that distributions would begin promptly.

U.S. District Judge William Alsup issued the preliminary approval in September and has since retired. He had ruled last summer that training AI on copyrighted books was not illegal, but that Anthropic had wrongfully acquired millions of copies via pirate sites.

Anthropic’s deputy general counsel Aparna Sridhar praised the ruling as a landmark, noting that training AI on books is fair use under copyright law. The settlement, the first major resolution in dozens of AI copyright lawsuits, signals a shift in how courts view AI training and authorship.