Anthropic has reportedly reached a preliminary settlement in a significant class action lawsuit initiated by notable authors, steering clear of a potential financial disaster if the case had reached a trial. The settlement is anticipated to be finalized on September 3, with specifics to be disclosed later.
The lawsuit began in 2024 when authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson accused Anthropic of unlawfully utilizing their works to train its artificial intelligence models. Earlier, a California district court judge, William Alsup, had issued a summary judgment mostly favoring Anthropic, declaring the usage of the books as “fair use” and therefore legal. However, the judge did indicate that some books were acquired via "shadow libraries," including the infamous LibGen, which constituted piracy. Thus, while the authors could not win on all fronts, they maintained the right to challenge Anthropic in a class-action context for the piracy claims.
Under U.S. copyright law, statutory damages for piracy begin at $750 per infringed item. Given that Anthropic was believed to hold about 7 million works, the stakes were monumental, with potential penalties climbing into the billions, and reports suggested possible damages exceeding $1 trillion.
Legal experts noted the unexpected nature of the settlement, especially as Anthropic had been vigorously contesting the lawsuits in court and had recently engaged a new legal defense team. Concerns remain that unresolved author sentiments could lead to discontent once the settlement terms are disclosed, especially since many authors were only beginning to learn of their status in the case.
Anthropic continues to face other copyright lawsuits, particularly a notable claim from major record labels including Universal Music Group, who allege that the company illegally trained its AI on copyrighted lyrics. Recent developments show these plaintiffs have amended their claims to assert that Anthropic employed peer-to-peer services like BitTorrent for illicit downloads.
While settlements do not establish legal precedents, the outcome of this case will likely attract attention as numerous other high-profile AI copyright cases navigate the judicial system.
