Bartz v. Anthropic and the Misframing of Fair Use Factor One Revealed Purpose, Transformative Character, and Expressive Competition in AI Training

Bartz v. Anthropic risks becoming a foundational misstep in generative AI copyright litigation, holding that training large language models on copyrighted books is "exceedingly transformative" under the first fair use factor. This Article argues that the decision misreads and misapplies § 107(1) by defining the "character" of the use at the level of intermediate mechanics-training and weight conversion-rather than anchoring the inquiry in the operative, market-facing purpose that motivates and monetizes the copying.