Anthropic Settles $1.5B Lawsuit Over AI Training on Pirated Books

Anthropic, the AI startup backed by Amazon and Alphabet, has agreed to pay $1.5 billion to settle a class-action lawsuit filed by U.S. authors. The plaintiffs alleged the company trained its Claude AI chatbot using pirated books without permission. Each of the 500,000 affected authors will receive approximately $3,000, and Anthropic will destroy illegally obtained book files.

The lawsuit, initiated in 2024 by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accused Anthropic of sourcing training data from shadow libraries like Library Genesis. While a 2025 federal ruling affirmed that training AI on legally acquired copyrighted material constitutes fair use, the court found Anthropic liable for using pirated content. The settlement avoids a potentially costly trial scheduled for December 2025.