Court Sanctions Anthropic’s Historic $1.5 Billion Resolution in AI Copyright Dispute

A federal judge in San Francisco has sanctioned a landmark $1.5 billion settlement concerning an AI copyright lawsuit initiated by a group of authors against Anthropic. The plaintiffs claimed that the company utilized unauthorized copies of their literary works to develop its large language model, Claude. Initially, Judge William Alsup, who presided over the case, determined that while some use of the authors’ works for training was deemed fair, Anthropic did violate copyright laws by storing 7 million pirated books in a central repository.

In September of the previous year, Alsup rejected an earlier settlement proposal, expressing concerns that it inadequately addressed the authors’ needs and was overly coercive. Following his retirement, the case was assigned to US District Judge Araceli Martinez-Olguin, who has now approved the settlement terms. Anthropic’s new settlement website allows authors to verify if their works are included in the compensation scheme, which encompasses over 480,000 titles.

Under the approved settlement, authors are projected to receive around $3,000 for each affected work, and Anthropic is mandated to eliminate the unauthorized copies from its holdings. Nevertheless, some authors feel the compensation is insufficient and have opted to withdraw from the agreement to pursue individual lawsuits against Anthropic.