US judge approves Anthropic's $1.5B copyright settlement

Final Approval Granted in San Francisco

A federal judge in San Francisco granted final approval to Anthropic's $1.5 billion copyright class-action settlement with authors on July 20, 2026, resolving allegations that the AI company used books to train its Claude chatbot without permission. U.S. District Judge Araceli Martinez-Olguin of the U.S. District Court for the Northern District of California issued the order in *Bartz v. Anthropic*, the Authors Guild confirmed.

The settlement caps one of the largest copyright disputes in the generative AI sector. Anthropic deputy general counsel Aparna Sridhar stated: "We reached this settlement in 2025, after the court's landmark ruling that training AI on books is fair use under copyright law — which remains the law today," according to Reuters.

Scope and Remedies

The settlement is narrowly scoped: it releases claims only for Anthropic's past acquisition and copying of works, not future conduct or output-based claims, the Authors Guild clarified. A concrete remedy requires Anthropic to destroy original files of works torrented or downloaded from Library Genesis and Pirate Library Mirror, subject to legal preservation obligations.

Attorney Fees and Distribution

Judge Martinez-Olguin awarded attorneys more than $101 million out of the $187.5 million requested in fees, according to Reuters. The judge rejected complaints about settlement size, stating they were "not grounded in a realistic assessment of the overall risks and rewards of a trial." Bloomberg Law reported that $122 million of the total deal would go to attorneys and litigation costs.

Significance for AI Training Liability

The settlement reflects ongoing litigation risk for large AI companies over training data sources, even where companies argue fair use protections apply. Anthropic's settlement follows similar disputes with authors and publishers across the generative AI industry. The final approval order marks the close of a major civil proceeding but does not establish precedent limiting future copyright claims against AI developers, as the court did not rule on the merits of the underlying allegations.