Federal Judge Approves $1.5B Anthropic Settlement Over Pirated Books Used to Train Claude
Key Takeaways
- ▸$1.5B settlement represents the largest known copyright recovery in history, with authors and publishers receiving approximately $3,000 per book
- ▸Over 91% of 482,000 affected authors and publishers claimed their share, indicating strong participation and satisfaction with the class-action settlement
- ▸Court ruled that training AI on copyrighted books itself is not illegal, but Anthropic wrongfully sourced millions of books through pirate websites, establishing key legal precedent
Summary
A federal judge has approved a landmark $1.5 billion copyright settlement in which AI company Anthropic will compensate thousands of authors for using pirated copies of their works to train its Claude chatbot. District Judge Araceli Martínez-Olguín ruled that the class-action settlement provides "meaningful relief" to affected authors and publishers, with approximately 91% of the more than 482,000 books covered having been claimed, yielding roughly $3,000 per book for rights holders.
The settlement stems from a lawsuit originally filed by bestselling thriller novelist Andrea Bartz and two other authors in 2024. While the court previously ruled that training AI models on copyrighted books does not constitute copyright infringement per se, Judge William Alsup found that Anthropic wrongfully acquired millions of books through pirate websites. Plaintiff attorney Justin Nelson described the settlement as "the largest known copyright recovery in history."
The case represents a critical precedent in AI regulation and creator compensation. Anthropic's deputy general counsel framed the court's fair-use finding as validation for AI training practices, while dozens of similar AI copyright lawsuits remain pending. This settlement may influence how the industry balances AI development with author and publisher rights going forward.
- First major settlement among dozens of pending AI copyright lawsuits, likely to shape how the industry sources and compensates creators for training data
