Judge Criticizes Anthropic's $1.5 Billion Settlement Over AI Training Data

U.S. District Judge William Alsup has raised substantial concerns regarding Anthropic's proposed $1.5 billion settlement with authors who alleged the company used pirated books to train its AI chatbot, Claude. The settlement aims to compensate authors approximately $3,000 per book for around 500,000 works.

During a recent hearing, Judge Alsup criticized the lack of detailed information in the settlement, particularly the absence of a comprehensive list of affected works and authors. He emphasized the need for clarity in the claims process and set deadlines for the parties to provide this information by September 15, with additional clarifications due by September 22.

The judge also expressed concerns about potential undue influence from major industry groups, such as the Authors Guild and the Association of American Publishers, on the settlement terms. He questioned whether the agreement adequately addresses the interests of all affected authors and publishers.

In response, leaders from the Authors Guild and the Association of American Publishers defended the settlement, arguing that it represents a fair resolution given the complexities of the case. They contended that the judge's concerns might stem from a misunderstanding of the publishing industry's intricacies and the challenges associated with managing a complex claims process.

The lawsuit, initiated by authors including Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, accused Anthropic of using millions of pirated books to train its AI models without permission. This case is among the first major legal challenges addressing the use of copyrighted material in AI training, setting a precedent for future disputes in the rapidly evolving field of artificial intelligence.