Federal Judge Finalizes $1.5 Billion Settlement with Anthropic Over Unauthorized Book Training Data

A San Francisco federal court has given final approval to a landmark $1.5 billion settlement requiring AI company Anthropic to compensate authors whose copyrighted books were illegally used to develop its Claude chatbot system. District Judge Araceli Martínez-Olguín issued the ruling Monday, determining that the agreement delivers “meaningful relief” to the thousands of writers and publishing companies affected by the unauthorized use of their intellectual property.

The settlement covers approximately 482,000 books, with roughly 91 percent already claimed by eligible authors and publishers who will each receive approximately $3,000 per title. Class action attorney Justin Nelson characterized the agreement as “the largest known copyright recovery in history,” pledging to distribute funds to affected parties without unnecessary delay.

U.S. District Judge William Alsup, now retired, had granted preliminary approval last September before handing down a complicated ruling the previous summer. That decision found that using copyrighted material to train artificial intelligence systems was not inherently unlawful, but concluded Anthropic had illegally obtained millions of books through piracy networks.

Bestselling author Andrea Bartz initiated the lawsuit alongside two fellow writers in 2024, launching the class action that ultimately resulted in this settlement.