Particle.news

Download on the App Store

Judge Finds Anthropic’s AI Training on Books Fair Use, Piracy Trial to Proceed

Deeming the use of purchased books transformative under copyright law, the court scheduled a December trial to determine damages for storing seven million pirated titles

Image
Image
Anthropic wins AI copyright ruling, judge says training on purchased books is fair use

Overview

  • U.S. District Judge William Alsup granted summary judgment that using copyrighted books to train Anthropic’s Claude model is “exceedingly transformative” and qualifies as fair use under Section 107 of the Copyright Act.
  • Alsup ruled that Anthropic’s creation of a permanent “central library” containing over seven million pirated book copies constituted copyright infringement and fell outside fair use protections.
  • A jury trial is set for December to assess statutory damages for the piracy claims, which may reach up to $150,000 per work even though later purchases of some titles could reduce the penalty.
  • The decision marks one of the first detailed applications of the fair use doctrine to generative AI and is expected to guide dozens of pending copyright lawsuits against AI developers.
  • Authors Andrea Bartz, Charles Graeber and Kirk Wallace Johnson, who filed the suit alleging large-scale theft of their works, are seeking class action status as the case moves toward trial.