Federal Judge Rejects $1.5B Anthropic AI Book Piracy Settlement

Anthropic recently agreed to a $1.5 billion settlement with authors whose works were allegedly pirated to train its AI chatbot Claude, but a federal judge has rejected the deal pending further clarification. U.S. District Judge William Alsup expressed concerns about the fairness of the claims process and transparency regarding the authors included in the class action, scheduling a follow-up hearing in late September to reassess the settlement. The judge also warned of potential conflicts involving the Authors Guild and the Association of American Publishers influencing authors without full consent. Should the settlement not be approved, the case could proceed to trial, with Anthropic potentially facing damages exceeding $10.5 trillion, which would be financially devastating. Meanwhile, a new lawsuit accuses Apple of similarly using pirated books to train its AI models, alleging that the company exploited unauthorized datasets without compensating authors, seeking damages up to $2.5 billion. These developments reflect a broader wave of legal challenges confronting AI companies over unauthorized use of copyrighted material for training their models.