Crypto news

22.07.2026
09:43

Anthropic's $1.5 Billion Settlement: Court Approves Largest Compensation for Book Piracy

A federal court in San Francisco has finally approved a $1.5 billion settlement agreement by the company Anthropic in a class-action lawsuit filed by authors who accused the AI developer of illegally using their books to train the Claude language model. This decision has become a landmark precedent in the field of copyright law and artificial intelligence.

District Judge Araceli Martínez-Olguín rejected objections from some plaintiffs who considered the payments insufficient and recognized the deal as lawful. According to expert estimates, this settlement is the largest known copyright case in U.S. history related to AI technology.

The lawsuit was filed in 2024. The authors claimed that Anthropic used pirated versions of their works—more than seven million books—without permission when training Claude. Earlier, Judge William Alsup ruled that training AI on copyrighted books could potentially fall under the fair use doctrine, but he found the storage of pirated copies in the company's "central library" to be a violation.

Anthropic's Deputy General Counsel Aparna Shridhar confirmed that the agreement was reached in 2025—after Alsup's ruling was issued. Notably, more than 91% of authors and publishers covered by the deal's terms have already filed claims to receive their share of the payments.

Lead attorney for the plaintiffs, Justin Nelson, called the agreement historic. However, some authors opposed it, disputing the total amount, the size of attorney fees, and the exclusion of certain rights holders from the agreement. The judge rejected these claims and approved a payment of $101 million to the lawyers out of the requested $187.5 million.

Individual authors and publishers have opted out of the class settlement and filed separate lawsuits against Anthropic. These proceedings are ongoing.

Expert opinion: The approval of this settlement sends a clear signal to the entire AI industry: using pirated content to train models will no longer go unpunished. However, the fact that 91% of rights holders have already agreed to the terms suggests that the market is ready for compromises. The key question remains open: where is the line between fair use and outright piracy? Future precedents, including lawsuits against ByteDance, will provide the final answer.