Judge approves $1.5B Anthropic authors settlement

A U.S. federal judge in San Francisco gave preliminary approval to a $1.5 billion settlement between AI firm Anthropic and authors who allege nearly half a million books were pirated to train the company's chatbots. The deal—backed by the Authors Guild and the Association of American Publishers—would pay authors and publishers about $3,000 per covered work, resolves claims about how materials were acquired, and follows a June ruling that found training can be transformative fair use but that Anthropic wrongfully obtained works from pirate sites.