Anthropic Receives Nearly 120,000 Author Claims in $1.5B Copyright Settlement

Gate News message, April 18 — Nearly 120,000 authors and copyright holders have filed claims to share in Anthropic's $1.5 billion class-action settlement over the unauthorized use of books in AI training, according to court filings in California. The claims cover 91% of more than 480,000 eligible works. Anthropic is scheduled for a May 14 hearing, where a judge will decide whether to grant final approval to what has been described as the largest settlement in a U.S. copyright case.

The lawsuit originated after authors alleged that Anthropic used pirated books to train Claude. In a June 2025 ruling, Judge William Alsup determined that training on legally obtained books was "quintessentially transformative" and constituted fair use. However, the judge found that downloading and storing more than 7 million pirated books from sites such as Library Genesis (LibGen) and Pirate Library Mirror (PiLiMi) to build a central library violated copyright holders' rights, even if those books were not necessarily used for AI training.

Settlement eligibility required titles to appear on the court-approved "Works List," meaning they were among the LibGen and PiLiMi files Anthropic downloaded and had been timely registered with the U.S. Copyright Office. This left approximately 482,460 eligible books out of the more than 7 million copies downloaded.

The $1.5 billion payout represents less than 1% of Anthropic's $183 billion valuation. Some observers view the settlement as a potential competitive advantage for well-funded AI firms, as smaller companies may struggle to manage similar litigation costs, and the case may accelerate industry adoption of licensed data over pirated sources.