Photo by Josh Withers on Pexels

A federal judge has temporarily blocked Anthropic’s proposed $1.5 billion settlement in a copyright infringement lawsuit, raising questions about whether the agreement adequately protects authors’ interests. Judge William Alsup voiced concerns that the terms of the settlement, intended to resolve claims that Anthropic used copyrighted books to train its AI models, might unduly favor class action lawyers over the authors themselves. The judge’s decision comes after he previously ruled that using legally purchased books for AI training constituted fair use. However, he acknowledged potential liability for Anthropic if illegally downloaded copyrighted material was used in the training process. Judge Alsup has requested more details regarding the settlement’s claims process and emphasized the importance of clear and comprehensive notification to all affected class members. While one industry group lamented the delay, arguing the settlement should resolve existing disputes rather than create new ones, the judge has scheduled a follow-up hearing for September 25th to further evaluate the agreement.