Attorneys contrast Anthropic and Meta fair-use rulings on market harm in AI book-training cases

Attorneys at McKool Smith compare Judge Alsup's ruling in Bartz v. Anthropic with Judge Chhabria's decision in Kadrey v. Meta. Both courts held AI training on copyrighted books could be fair use, but diverged on pirated-material handling (Alsup treated Anthropic's general-purpose pirated-book library as infringement; Chhabria treated Meta's use of shadow-library copies as part of a transformative process) and on market dilution: Alsup dismissed fears of competition from AI-generated substitutes, while Chhabria called the risk unprecedented and warned future plaintiffs could succeed with a better-developed record of market harm.

The article notes the U.S. Copyright Office's pre-publication 'Copyright and Artificial Intelligence, Part 3' report expressed concerns similar to Chhabria's, and that Anthropic filed a motion for interlocutory appeal of Bartz to the Ninth Circuit on July 14.