Judge approves a $1.5B Anthropic settlement over pirated books used to train the Claude chatbot - AP News

What the report says

The Associated Press reported that a federal judge has approved a $1.5 billion copyright settlement resolving claims that artificial intelligence company Anthropic used pirated books to train its Claude chatbot. The ruling, issued Monday by U.S. District Judge Araceli Martínez-Olguín in San Francisco, clears the way for payments to authors and publishers covered by the class action, with compensation expected to be about $3,000 per book.

According to AP, the settlement covers more than 482,000 books, and rights holders have submitted claims for roughly 91% of them. Martínez-Olguín described the agreement as providing meaningful relief. Plaintiff attorney Justin Nelson called it the largest known copyright recovery, and said the class is expected to receive distributions as soon as possible.

The case was initially brought in 2024 by bestselling thriller writer Andrea Bartz and two other authors. AP reported that retired U.S. District Judge William Alsup had previously given preliminary approval to the deal and issued a mixed ruling: he found that training AI systems on copyrighted books was not unlawful in itself, but that Anthropic improperly obtained millions of books from pirate websites.

Anthropic deputy general counsel Aparna Sridhar pointed to that earlier ruling as significant for the company’s position that AI training on books can qualify as fair use. The settlement matters because it is the first major resolution among many pending lawsuits testing how copyright law applies to AI training data, a central legal question for publishers, authors and technology firms.