Judge Lambastes Proposed $1.5 Billion AI Copyright Deal, Citing Unfairness to Authors

In a significant legal development sending ripples through the burgeoning artificial intelligence industry, a federal judge has cast a critical eye on a proposed settlement involving AI powerhouse Anthropic and a collective of authors. U.S. District Judge Dolly M. Gee of Los Angeles did not mince words, raising profound skepticism over the terms of a deal designed to resolve claims that Anthropic unlawfully used copyrighted books to train its advanced AI chatbots.

The core of the dispute revolves around the accusation that Anthropic, like other leading AI firms, ingested vast quantities of literary works without permission or compensation, utilizing them to fuel the algorithms behind its generative AI models.

The proposed settlement in question was between Anthropic and several individual authors, alongside the Authors Guild, a prominent advocacy group that is also part of a larger lawsuit against Anthropic, OpenAI, and Stability AI.

Judge Gee's primary concern centers on the extraordinary breadth of the proposed agreement.

She noted that the settlement, if approved, would grant Anthropic a release from future claims that she estimated could be worth "up to $1.5 billion" to the company. This substantial sum, the judge suggested, might come at an unfair cost to a vast number of authors not directly party to the settlement, particularly those who are not members of the Authors Guild, whose intellectual property rights could be inadvertently undermined.

During a recent hearing, Judge Gee highlighted the potential for "unjust enrichment" for Anthropic.

She pointed out that the settlement's language was alarmingly broad, stipulating a release from claims "worldwide" for "any written work or other content" owned by class members, even if those works were not specifically involved in the initial lawsuit. This sweeping release, according to the judge, could effectively insulate Anthropic from future accountability for using copyrighted material from countless authors without proper licensing or payment.

The judge questioned how the settlement could ethically release claims from authors who are not explicitly represented or compensated through this particular agreement.

This legal challenge underscores a growing tension between AI developers' need for massive datasets to train their models and the fundamental rights of creators to control and profit from their original works. The outcome of this and similar cases could set crucial precedents for how intellectual property is handled in the age of generative AI.

This scrutiny is part of a larger trend, as AI companies find themselves embroiled in a rapidly expanding web of lawsuits from authors, artists, and news organizations, all seeking to protect their creative output from unauthorized use by AI systems.

Judge Gee has given the involved parties until January 10 to either respond to her concerns or submit an amended settlement proposal, indicating that the path to a resolution in this high-stakes copyright battle is far from clear.