Recent rulings in the cases of Kadrey v. Meta and Bartz v. Anthropic have highlighted the divergent judicial perspectives on whether training AI models on copyrighted material constitutes fair use. In San Francisco, U.S. District Judge Vince Chhabria granted Meta’s motion for summary judgment in Kadrey v. Meta, essentially dismissing the authors’ claims. However, he noted that this decision was based on ineffective legal arguments presented by the plaintiffs and did not affirm the legality of Meta’s use of the copyrighted works (Law.com).
Conversely, U.S. District Judge William Alsup in Bartz v. Anthropic found that training AI on legally obtained books was “spectacularly transformative” and fell under fair use. However, the use of pirated books raised infringement questions warranting further trial consideration. Commentators like Jason L. Haas argue these cases offer more guidance for plaintiffs than a victory for tech firms, signaling that acquiring content lawfully might be crucial for considering fair use defenses (The Recorder). For further details on these rulings, visit the original article.