Class Action Against Anthropic: Authors Allege Copyright Infringement from Pirated Works
U.S. District Judge allows class action against Anthropic for alleged copyright infringement involving pirated books used in AI training.
Key Points
- • Judge Alsup permits a class action lawsuit against Anthropic for copyright infringement.
- • Authors claim Anthropic downloaded up to 7 million pirated works for AI training.
- • Total damages could exceed one trillion dollars if all claims are validated.
- • Anthropic considers appealing the ruling due to challenges in proving ownership.
In a significant development for the AI landscape, U.S. District Judge William Alsup ruled on July 18, 2025, permitting a class action lawsuit against Anthropic on behalf of several authors. The authors allege that the company improperly used millions of pirated books to train its AI model, Claude, leading to a potential financial reckoning that could amount to trillions in damages.
The lawsuit highlights claims that Anthropic downloaded works from unidentified 'pirate libraries', such as LibGen and PiLiMi, resulting in a collection estimated to comprise up to 7 million books. This massive infringement could trigger statutory damages of up to $150,000 per book, positioning the total potential compensation to exceed one trillion dollars, should all infringements be substantiated (Research Item ID: 14899).
Judge Alsup noted the straightforward nature of proving damages due to the existing metadata associated with pirated works. In contrast, Anthropic has raised concerns regarding the challenge of identifying rights holders for all involved works, arguing that this complexity was not thoroughly addressed during the proceedings. However, the judge dismissed this argument, confirming that the case exemplifies the characteristics conducive to class action status, given that multiple authors have reportedly suffered identical harm (Research Item ID: 14905).
The plaintiffs in the case—Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson—assert that their copyrighted materials were used without their consent or any form of remuneration. They emphasize that Anthropic's actions violate the Copyright Act through the unauthorized storage of their works in a centralized manner, notwithstanding a previous ruling in which Anthropic's training practices could potentially be considered 'fair use' concerning legitimately sourced materials (Research Item ID: 14905).
This class action exacerbates the ongoing legal tensions between AI companies and authors over copyright issues, highlighting a pivotal moment in the evolving narrative of AI's relationship with creative content. As the case progresses, its implications may set crucial precedents, influencing how AI companies negotiate licensing with creators and copyright holders.
With the potential for colossal damages looming, Anthropic is reportedly considering appealing Judge Alsup’s ruling, indicating that this high-stakes legal battle is far from resolution and could shape the future landscape of AI training practices.