Judge questions fairness of Anthropic's $1.5 billion AI training data settlement with authors

A judge refused to approve Anthropic's $1.5 billion settlement with authors over AI training data piracy. US District Judge William Alsup expressed concerns that the settlement was incomplete and potentially unfair to authors, citing unresolved issues regarding claimant lists and payment allocation. The judge has ordered Anthropic and authors' lawyers to revise the settlement, with a new preliminary approval decision expected by October 10.