Recently, a federal court in California granted final approval to the $1.5 billion copyright settlement reached between AI company Anthropic and a group of authors. This ruling brings to a close the class action lawsuit filed in 2024 alleging that Anthropic used copyrighted works without authorization to train its large language model, Claude.

On July 21, 2026, U.S. District Judge Araceli Martínez-Olguín issued a final approval order, finding that the settlement agreement was fair and reasonable, free from collusion, and provided substantial relief to certified class members in the form of $1.5 billion (approximately 10.15 billion RMB) in copyright royalties. The judge also approved approximately $101.5 million in lead counsel fees for the class action, as well as approximately $2.6 million in expense reimbursements and an $18.2 million contingency reserve, and awarded each of the three class representatives a $15,000 incentive payment.

The settlement stems from a lawsuit filed by a group of authors in 2024, alleging that Anthropic committed “mass theft” by downloading copies of known pirated works and inputting them into large language models for training. The case’s original presiding judge, William Alsup, had previously granted Anthropic a partial summary judgment, ruling that its use of legally purchased books to train AI constituted fair use, but ordered the company to stand trial on charges related to the storage of millions of pirated books. When Judge Alsup preliminarily approved the agreement in September 2025, he called it “the largest copyright class-action settlement in history.”

In the final approval order, Judge Martínez-Olguín dismissed various objections to the settlement agreement, including those challenging its scope and arguing that the $1.5 billion amount was too low. The judge noted that notices had been sent to 506,194 potential class members (representing 99.5% of the works on the list in question) with no returns, and that the notice rate fell within the range deemed reasonable by the Federal Judicial Center. Regarding the challenge to the settlement amount, the judge stated that the objection “was not based on a realistic assessment of the overall risks and rewards of a trial” and emphasized that “even if the trial were won, post-trial proceedings could jeopardize class compensation and result in significant delays.”

Aparna Sridhar, Deputy General Counsel at Anthropic, stated that the company reached the settlement in 2025 following the court’s ruling that “training AI using books constitutes fair use,” and expressed satisfaction that over 91% of authors and publishers have already claimed their payments. Rachel Geman, co-lead counsel for the class action, said the settlement “upholds the rights of authors and publishers in the age of artificial intelligence,” while co-lead counsel Justin A. Nelson called the agreement “historic” and stated that funds would be distributed to class members as soon as possible.

Scan the QR code to view the full text of the ruling