Anthropic Settles AI Copyright Lawsuit with Book Authors

Anthropic has reached a preliminary settlement in a major class action lawsuit brought by a group of authors, resolving one of the most closely watched legal disputes over AI copyright infringement in the industry. This development allows the AI company to sidestep a potentially catastrophic financial outcome that could have reached into the billions of dollars.

According to a recent legal filing, the settlement is expected to be finalized by September 3, though specific terms remain undisclosed. Attorneys for the plaintiffs have not yet commented publicly, and Anthropic has also declined to provide additional details at this stage.

The lawsuit originated in 2024 when authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson accused Anthropic of unlawfully using their copyrighted books to train its AI models. In a significant ruling this past June, California District Court Judge William Alsup issued a summary judgment that largely favored Anthropic, concluding that the company’s use of the material fell under fair use protections.

However, the judge also determined that Anthropic had obtained some of the works through shadow libraries, including the controversial site LibGen, which he classified as piracy. This left the door open for the authors to pursue a class action trial focused specifically on these unauthorized acquisition methods, a proceeding that had been scheduled for December.

Under U.S. copyright law, statutory penalties for piracy begin at $750 per infringed work. Given that Anthropic’s training library is believed to include around 7 million works, the company faced potential damages that could have soared into the trillions of dollars, a risk that made settlement a pragmatic choice.

Edward Lee, a law professor at Santa Clara University who specializes in technology and intellectual property, described the situation as a dramatic shift. “Anthropic had been vigorously defending itself across multiple fronts,” he noted. “But after the judge’s ruling, their position at trial was extremely precarious. The threat of astronomical statutory damages clearly influenced their decision to settle.”

Many authors who might have been part of the class action were only recently notified of their eligibility. The Authors Guild, a leading organization for professional writers, began alerting members earlier this month, and a formal list of affected works was due to be submitted to the court on September 1. This timing means that numerous writers were not directly involved in the settlement discussions.

James Grimmelmann, a digital law professor at Cornell, pointed out that the real test will come when the settlement terms are made public. “We’ll have to see how authors react,” he said. “Their response will serve as a key indicator of where copyright holders stand on these issues more broadly.”

Anthropic continues to face additional legal challenges related to copyright. Among the most prominent is a lawsuit from several major record labels, including Universal Music Group, which alleges that the company trained its AI on copyrighted song lyrics without permission. Recent court filings suggest the plaintiffs are now also accusing Anthropic of using BitTorrent to illegally download music.

While settlements do not establish legal precedent, the resolution of this case will undoubtedly be analyzed closely as numerous other high-stakes AI copyright lawsuits continue to make their way through the judicial system.

(Source: Wired)