Inside the Largest Copyright Recovery in History

John is joined by Rachel Geman, partner at Lieff Cabraser Heimann & Bernstein, LLP, and Justin A. Nelson, and Rohit Nath, both partners at Susman Godfrey. They discuss the groundbreaking $1.5 billion copyright class action settlement Rachel, Justin, and Rohit reached with AI company Anthropic on behalf of the authors of copyrighted materials —the largest copyright recovery in history.

The case involved Anthropic’s use of over 450,000 copyrighted works—mostly books—sourced from pirated sites like Library Genesis and Z-Library. These works were used to train large language models (LLMs). The case centered on infringing conduct stemming from the download and use of pirated copies of copyrighted works. Judge William Alsup, who presided over the case, found that Anthropic’s downloading of pirated works was “irredeemably wrong” and constituted infringement. He also ruled that using legitimately obtained books to train AI was transformative and, therefore, fair use—a finding the plaintiffs disagreed with. A trial was scheduled but avoided when the parties reached a $1.5 billion settlement shortly after fact discovery closed.

The settlement compensates authors and publishers at an average rate of approximately $3,000 per work. The settlement also reflects contractual author-publisher splits and employs a structured claims process overseen by a special master. Under the agreement, Anthropic must also destroy the infringing copies and certify they were not used to train its commercial models. This resolution, the largest known copyright recovery to date, was approved after detailed scrutiny of its fairness and administration.

Podcast Link: Law-disrupted.fm
Host: John B. Quinn
Producer: Alexis Hyde
Music and Editing by: Alexander Rossi