Anthropic Resolves Prominent AI Copyright Case Filed by Authors

Anthropic has achieved a preliminary settlement regarding a class action lawsuit filed by several well-known authors, signifying a pivotal moment in one of the most important AI copyright lawsuits in history. This development enables Anthropic to circumvent what could have been a financially ruinous court verdict.

The settlement is set to be finalized on September 3, with additional details anticipated, according to a legal filing made public on Tuesday. Anthropic has opted not to comment. “This groundbreaking settlement will benefit all class members. We are eager to disclose the details in the upcoming weeks,” stated Justin Nelson, an attorney representing the plaintiffs, in a statement to WIRED.

In 2024, three authors—Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson—filed a lawsuit against Anthropic, claiming that the startup unlawfully utilized their works to train its AI models. In June, California district court judge William Alsup delivered a summary judgment in Bartz v. Anthropic that largely favored Anthropic, ruling that the company’s use of the books constituted “fair use” and was thus lawful.

However, the judge determined that the way Anthropic acquired some of the content, specifically through downloading from shadow libraries, including a well-known site called LibGen, amounted to piracy. Alsup ruled that the authors could still pursue a class action against Anthropic for the alleged piracy; the legal conflict was scheduled to commence in December.

According to US copyright law, statutory damages for this type of piracy begin at $750 per infringed work. Given that Anthropic’s library was estimated to hold about 7 million works, the AI firm was potentially facing court-imposed penalties reaching billions, possibly exceeding $1 trillion.

“This is a remarkable turn of events, particularly considering how aggressively Anthropic was contesting this case in two courts. Moreover, the company recently appointed a new trial team,” remarks Edward Lee, a law professor at Santa Clara University who closely monitors AI copyright litigation. “However, their defenses at trial were minimal, especially in light of Judge Alsup’s ruling. Thus, Anthropic was facing the possibility of statutory damages in ‘doomsday’ ranges.”

Most authors eligible for the class action were just beginning to receive notifications about their participation. Earlier this month, the Authors Guild, a trade organization for professional writers, alerted authors about their eligibility, while attorneys for the plaintiffs were expected to present a “list of affected works” to the court on September 1. This indicates that many authors were not involved in the negotiation process.

“The crucial question is whether there will be considerable dissent from within the author community once the settlement terms are revealed,” comments James Grimmelmann, a professor specializing in digital and internet law at Cornell University. “That will serve as a significant indicator of the sentiments among copyright owners.”

Anthropic continues to contend with several other copyright-related legal challenges. One prominent case involves a coalition of major record labels, including Universal Music Group, which assert that the company unlawfully trained its AI systems on copyrighted lyrics. The plaintiffs have recently filed to amend their case, claiming that Anthropic utilized the peer-to-peer file-sharing service BitTorrent for illegal song downloads.

Settlements do not establish legal precedent, but the particulars of this case are likely to be closely monitored as numerous other notable AI copyright cases progress through the judiciary.

Update: 8/26/25, 11:40 pm EST: This story has been revised to include comments from an attorney representing the plaintiffs.