The ruling authorizes roughly $3,000 per work across about 500,000 books and underscores that the payout resolves Anthropic’s piracy exposure without creating binding precedent on whether AI training is fair use.
A U.S. federal judge approved Anthropic’s $1.5 billion class action settlement with authors and publishers over pirated books used in connection with training Claude, finalizing what plaintiffs called the largest known copyright recovery in history. U.S. District Judge Araceli Martínez-Olguín found the deal “fair and adequate,” rejected objections that it undervalued claims, and cleared distributions of roughly $3,000 per work across approximately 500,000 titles, with more than 92% of eligible rightsholders submitting claims. Anthropic also agreed to destroy the pirated files it had accumulated. The settlement follows a split ruling issued in June 2025 by then-presiding Judge William Alsup, who found that training an AI model on copyrighted books was transformative fair use under 17 U.S.C. § 107, but held that Anthropic’s maintenance of a permanent central library of more than seven million pirated books was not protected. That distinction drove Anthropic’s legal exposure because willful infringement can carry statutory damages of up to $150,000 per work under 17 U.S.C. § 504(c)(2). Anthropic settled before a December 2025 trial on damages. The case, Bartz et al. v. Anthropic PBC, does not establish binding law for the wider AI industry because the fair-use ruling was a district court decision that will not be reviewed on appeal. That leaves other AI copyright suits against Google, Meta, OpenAI and Midjourney free to reach different conclusions, even as the Anthropic deal may serve as an early market benchmark for AI use of copyrighted books.