Authors Hail “Historic” Settlement in Anthropic Class Action
▼ Summary
– Authors and Anthropic have reached a “historic” settlement in a class-action lawsuit over AI training data, with a motion for preliminary approval due by September 5.
– The class action was certified by Judge William Alsup as potentially the largest copyright class action ever, allowing up to 7 million claimants to join.
– The lawsuit was initiated by three authors but expanded due to the alleged illegal downloading of millions of books to train Anthropic’s AI models.
– Industry advocates warned that if all authors filed claims, it could financially ruin the entire AI industry, and Anthropic previously argued the lawsuit threatened its survival.
– Settlement terms are binding but undisclosed, with authors’ lawyer Justin A. Nelson confirming it will benefit all class members and details will be revealed soon.
A major class action lawsuit against artificial intelligence company Anthropic appears to be nearing a landmark resolution, with authors and the firm reaching a settlement that could set a precedent for how creative works are used in AI development. The agreement, described as “historic” by legal representatives, addresses claims that the company improperly used copyrighted materials to train its AI models without permission or compensation.
US District Judge William Alsup confirmed this week that both parties have reached a settlement in principle, with a formal motion for preliminary approval expected by early September. This development follows the judge’s decision to certify what has been called the largest copyright class action in history, a move that drew sharp criticism from AI industry advocates who warned that the scale of potential claims could destabilize the entire sector.
Although the case was originally brought by just three authors, Andrea Bartz, Kirk Wallace Johnson, and Charles Graeber, the class was expanded to include as many as seven million claimants. This broad certification was based on allegations that Anthropic downloaded a vast number of books without authorization to train its AI systems, a practice the plaintiffs argued violated their intellectual property rights.
The exact terms of the settlement remain confidential, and it is unclear whether the expansive class certification directly influenced Anthropic’s decision to settle. What is known, however, is that the agreement is binding, and attorneys for the authors have characterized it as a significant victory. Justin A. Nelson, a lawyer representing the plaintiffs, stated that the settlement will benefit all class members and promised further details in the coming weeks.
Anthropic, founded by former OpenAI employees in 2021, had previously argued that the lawsuit posed an existential threat to the company and the broader AI industry. The firm warned that a flood of claims could prove financially devastating, though it has not yet commented publicly on the settlement. The outcome of this case is being closely watched, as it may influence how other AI firms approach training data sourcing and copyright compliance in the future.
(Source: Ars Technica)
