Both cases turn on a single question: whether AI training on copyrighted material constitutes fair use or infringement. The plaintiffs argue that Google and Meta improperly accessed textbooks, scientific papers, novels, and journal articles to develop their systems. The legal standard remains unsettled. A June 2025 ruling in Bartz v. Anthropic found that training on legally acquired books qualified as fair use, and courts similarly favored Meta in a separate decision. However, judges are increasingly applying a fact-specific market-harm test rather than issuing blanket rulings. The emerging pattern suggests that training on pirated or unlicensed content creates near-certain liability, while properly licensed data strengthens fair use arguments.
Attorneys should monitor these cases closely. Over 35 AI training-data lawsuits are now active across major Western jurisdictions, with potential damages in the largest cases reaching billions of dollars. The outcomes will determine whether AI companies must shift to extensive licensing regimes or can continue relying on fair use defenses for unlicensed data. 2026 is shaping up as the year courts deliver clarifying decisions that will define the future of AI development and copyright law in the United States.