Greetings from Las Vegas. I'm here to do a keynote about AI for eBay Open 25. I did a bit of reading on the plane and, while I'm not a lawyer, you don't need to be one to know that this case is really important.
On July 17, 2025, U.S. District Judge William Alsup ruled that three authors (Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson) can represent all U.S. writers whose works Anthropic allegedly downloaded from pirate libraries LibGen and PiLiMi. According to Reuters, Alsup said Anthropic may have illegally downloaded as many as seven million books, exposing the Amazon and Alphabet‑backed startup to over a trillion dollars in damages.
This class certification is staggering. Industry groups warn that because statutory damages for willful infringement can reach $150,000 per work, a class involving up to seven million potential claimants would devastate the entire AI industry. Last week, Anthropic petitioned the Ninth Circuit to block what they call the largest copyright class action ever certified.
The Legal Landscape
Judge Alsup’s June 23 fair‑use ruling draws a crucial distinction. He held that training AI models on copyrighted books is fair use, calling the process “transformative, spectacularly so.” However, Anthropic’s decision to build a central library using 7 million unauthorized copies that included “pirated” books as well as scanned lawful copies went beyond fair use. He granted summary judgment for the training use but denied it for the pirated library, allowing damages claims to proceed.
Anthropic asked Alsup to allow an immediate appeal, arguing that he failed to perform the rigorous analysis required for class certification and instead relied on his “50 years” of courtroom experience. Alsup declined, holding that the Ninth Circuit should review the case only after the December 1, 2025 jury trial. That trial will determine how much Anthropic owes for building its pirate library.
Why the AI Industry Is Panicking
Anthropic’s petition argues that it now faces hundreds of billions of dollars in potential liability because the class certification was rushed at “warp speed” and could encompass seven million potential claimants. If each claimant sought the $150,000 statutory maximum, the damages would dwarf the revenues of most AI startups. This threat isn’t limited to Anthropic. Virtually every generative‑AI company has trained on publicly available data, so a precedent requiring licenses for each copyrighted work would force model builders to negotiate millions of individual deals or risk crushing legal exposure.
Industry groups including the Consumer Technology Association and the Computer and Communications Industry Association echo this alarm. In a recent amicus brief, they warned that the class certification “would threaten immense harm not only to a single AI company, but to the entire fledgling AI industry and to America’s global technological competitiveness.” They argue that the prospect of enormous liability exerts coercive settlement pressure that could stifle investment and innovation.
Strategic Implications
We should view this case as a test of how U.S. law will treat AI training. Today’s ruling says using copyrighted books to train large language models is legal, but retaining pirated copies for other purposes is not. That nuance may force AI developers to tighten data‑provenance controls and purge training libraries of unauthorized works. It may also accelerate moves to license content or develop synthetic data to mitigate litigation risk.
If the Ninth Circuit upholds the class certification, expect a wave of similar lawsuits against other model builders. Companies should prepare for more rigorous data‑governance audits and begin scenario‑planning for potential statutory damages. Likewise, policymakers may need to consider new statutory frameworks (such as a compulsory licensing scheme or a tailored fair‑use doctrine for AI) to balance innovation with copyright holders’ rights.
Read the Court Order
For those who want to read the ruling that ignited this controversy, the court’s June 23, 2025 order in Bartz v. Anthropic (Northern District of California, No. 3:24‑cv‑05417) is available here: Order Granting in Part and Denying in Part Anthropic’s Motion for Summary Judgment (PDF). It explains why Judge Alsup believes training is fair use while maintaining a pirated library is not.
I think I got all my facts straight. I know you'll correct me if I don't. As always your thoughts and comments are both welcome and encouraged. -s
As always, your thoughts and comments are both welcome and encouraged. -s
Shelly Palmer is the Professor of Advanced Media in Residence at Syracuse University’s S.I. Newhouse School of Public Communications and CEO of The Palmer Group, a consulting practice that helps Fortune 500 companies with technology, media and marketing. Named LinkedIn’s “Top Voice in Technology,” he covers tech and business for Good Day New York, is a regular commentator on CNN and writes a popular daily business blog. He's a bestselling author, and the creator of the popular, free online course, Generative AI for Execs. Follow @shellypalmer or visit shellypalmer.com.
