Judge Approves Anthropic’s $1.5 Billion Copyright Settlement Over Pirated AI Training Books

A U.S. judge has approved Anthropic’s $1.5 billion settlement in a landmark copyright case over allegations that the company used pirated books to train its AI models without permission. The agreement resolves claims by authors and publishers and is expected to pay eligible rightsholders about $3,000 per work, making it one of the largest copyright recoveries ever reported. The case matters because it sets a major precedent for how AI companies may be held financially accountable for using copyrighted material in training data. Anthropic has also agreed to destroy the pirated book datasets identified in the case and to compensate class members through a court-supervised process.