Federal judge approves Anthropic's $1.5B copyright settlement

U.S. District Judge Araceli Martínez-Olguín of the U.S. District Court for the Northern District of California approved Anthropic's $1.5 billion settlement on July 20, 2026, resolving a copyright class action brought by authors including Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who alleged the company used their books to train the Claude chatbot without permission Reuters.

Settlement Scope and Distribution

The settlement covers approximately 480,000 to 500,000 qualifying works and distributes roughly $3,000 per work to class members Reuters. Judge Martínez-Olguín characterized the settlement as the "largest known settlement of a U.S. copyright case" Reuters. The company must fund the settlement in installments through September 2027 and is required to destroy the pirated books it downloaded from sources including Library Genesis and Pirate Library Mirror Reuters.

Lawyer Fees and Court Decision

The judge awarded class counsel approximately $101.6 million in fees, rejecting the original request for $187.5 million. In her ruling, Judge Martínez-Olguín stated that a percentage-based fee structure would have produced a "windfall" for lawyers, signaling strict scrutiny of legal compensation in the high-value settlement Reuters.

Scope of Released Claims

The settlement resolves only past claims related to Anthropic's acquisition and copying of books used as AI-training inputs through August 25, 2025. Class members release claims from that period but retain the right to pursue future disputes. The company's documented use of pirated books from underground libraries became a central issue in the litigation, establishing a concrete factual basis for the damages award.

This settlement represents a significant moment in the AI training data liability landscape, establishing that companies face substantial financial exposure for unauthorized use of copyrighted works in large-scale model training operations. The case creates precedent for authors and publishers seeking compensation from AI companies, and similar litigation from other rights holders remains ongoing.