A US federal judge on Thursday paused the final approval of artificial intelligence company Anthropic’s proposed $1.5 billion settlement with authors, demanding more clarity on the agreement and signaling increased legal scrutiny for AI developers over training data. The case centers on accusations that Anthropic used pirated books to train its Claude chatbot.
U.S. District Judge Araceli Martinez-Olguin in San Francisco did not grant final approval, instead asking for more detail on issues including lawyers’ fees and payments to lead plaintiffs. The settlement had been initially greenlit by now-retired Judge William Alsup last September.
The class-action lawsuit, first filed in 2024, covers over 480,000 works. While Judge Alsup ruled last June that Anthropic's use of the works for training constituted "fair use," he found the company violated copyright by saving more than 7 million pirated books to a "central library." A trial to determine financial damages for that infringement was scheduled for December.
The outcome of this case is being closely watched as it could set a critical legal precedent for intellectual property rights in the generative AI era. A substantial penalty or a settlement that heavily favors authors could significantly raise operating costs and legal risks for AI developers like OpenAI and Google, potentially cooling investor sentiment in a sector that has attracted billions in capital. The delay also highlights dissent among authors, with some, including novelist Dave Eggers, opting out to file separate lawsuits against Anthropic.
This article is for informational purposes only and does not constitute investment advice.
