Judge approves Anthropic $1.5bn AI copyright settlement

A federal judge in San Francisco has granted final approval to a landmark $1.5 billion settlement between artificial intelligence company Anthropic and a class of authors, resolving a lawsuit alleging the company trained its Claude chatbot on pirated books. The settlement, approved on July 21, 2026, by U.S. District Judge Araceli Martinez-Olguin, is the largest known payout in a U.S. copyright case.

The class action lawsuit, filed in 2024 by novelists Andrea Bartz and Charles Graeber and nonfiction writer Kirk Wallace Johnson, alleged Anthropic improperly used copyrighted books to train its AI models. The case centered on the company's use of shadow library websites, specifically LibGen and Pirate Library Mirror, to obtain millions of book titles. The settlement covers an estimated 480,000 to 500,000 works, with rightsholders receiving approximately $3,000 per book, a figure the judge noted is four times the statutory minimum for copyright infringement.

The legal path to the settlement involved a critical distinction made by now-retired Judge William Alsup in a June 2025 ruling. Judge Alsup found that the act of training an AI model on lawfully purchased books constituted fair...