Artificial Intelligence

Judge Certifies Class Action Against Anthropic Over Pirated Books

Published JUL 30 2025

Story

Above: Dario Amodei, co-founder and CEO of Anthropic, participated in a panel session at the World Economic Forum in Davos, Switzerland, on Jan. 23, 2025. Image credit: Stefan Wermuth/Bloomberg/Getty Images

The Facts

Federal Judge William Alsup recently certified a class action lawsuit against Anthropic, allowing three authors to represent nearly every U.S. book author whose works were allegedly downloaded from pirate libraries LibGen and PiLiMi to train the company's Claude AI model.
Obsolete
CO/AI
Court filings reveal Anthropic downloaded approximately five million books from LibGen and two million books from PiLiMi between 2021-2022, with the company acknowledging the books covered likely number "in the millions."
Obsolete
JURIST
Judge Alsup previously ruled that training AI models on lawfully acquired books constitutes fair use, but found that Anthropic's wholesale downloading and storage of pirated books from shadow libraries was not protected by fair use.
Obsolete
Fortune
Statutory damages under copyright law range from $750 to $150,000 per work, with potential liability reaching $1.5 billion at minimum damages or $750 billion at maximum damages if five million works are covered.
Obsolete
CO/AI
The class certification covers actual and beneficial copyright owners of books with ISBN or ASIN numbers downloaded from LibGen and PiLiMi, but excludes Books3 due to insufficient metadata for identifying titles and authors.
PublishersWeekly.com
Beneschlaw
Anthropic filed a motion to stay the case and plans interlocutory appeals, while the trial is scheduled for Dec. 1, marking the first certified class action against an AI company over copyrighted training materials.
ChatGPT is Eating the World
Bloomberg Law News

Sources Split

LEFT

CENTER

RIGHT

ESTAB
CRIT

CENTER

PRO
ESTAB

The Spin

Narrative A

Anthropic adheres to values of fair use and transparency. Negative legal cases not only could inhibit Anthropic but also innovation in AI as a sector as a whole, which is vital to U.S. security.

InfoQ

White & Case LLP

Narrative B

The certification finally holds AI companies accountable for systematic copyright theft. Anthropic deliberately downloaded millions of pirated books and celebrated finding new sources of stolen content, treating authors' life's work as free training data. This landmark case ensures creators receive compensation for their intellectual property being used to build billion-dollar AI systems.

Beneschlaw

PublishersWeekly.com