Source: American Bazaar Online

A U.S. federal court approved a $1.5 billion settlement in a copyright dispute between authors and publishers and Anthropic, resolving allegations that the company misused books to train its Claude AI chatbot. The settlement, approved on July 21, 2026 by Judge Araceli Martinez-Olgun, is described as the largest known U.S. copyright settlement and the first major AI copyright dispute resolved before trial; preliminary approval had been granted by Judge William Alsup in 2025. Affected authors and publishers are to receive $3,000 per work for roughly 500,000 works, the court approved over $101 million in legal fees, and the settlement administrator reported that more than 91% of the covered authors and publishers claimed the payment.

The court found that Anthropic acquired books both by legally purchasing and scanning copies and by downloading works from piracy sites, including Library Genesis and Pirate Library Mirror, and concluded that storing more than seven million pirated books in a central repository violated copyright law. The ruling reiterated that prior judicial findings had suggested training an AI model on copyrighted text could qualify as fair use but emphasized that acquisition methods remain central to disputes. Anthropic agreed to the settlement to avoid trial and potential additional damages; Anthropic’s Deputy General Counsel Aparna Sridhar said the company reached the settlement in 2025 after the court’s landmark ruling on fair use. The piece notes that some authors opted out and filed separate lawsuits and that related litigation targets other major companies over AI training on copyrighted works.