The fair use doctrine underpinning copyright law was built for single-work analysis, and AI model training at scale has produced a category of disputes that framework was never designed to handle. J.S. Held, a global consulting firm, released its Q2 2026 AI Disputes Monitor on July 15, documenting copyright cases as the dominant category in its proprietary AI disputes dashboard, with regulatory and product liability challenges registered as emerging alongside them.
Copyright at the center
Fair use evaluates a secondary use across four factors: purpose, nature, the portion copied, and market effect. Applying that test to the training of a large-scale AI model asks courts to extrapolate from a doctrine built for one work at a time. Content-creator litigation is growing more sophisticated, the Q2 monitor finds, and that shift matters procedurally as well as substantively. The cases now before courts are pressing judges to address how AI-specific regulation intersects with intellectual property doctrine, building precedent in a space where neither the copyright statute nor the emerging regulatory framework supplies clear guidance.
Product liability enters the dispute stack
Product liability represents a meaningful extension of the categories the monitor tracks. Copyright suits tend to concentrate on the training phase. Product liability theories address the output: what the system generated, who relied on it, and what harm resulted. The Q2 2026 report captures this broadening as regulatory challenges appear alongside the liability claims, suggesting the dispute categories are moving beyond intellectual property as the primary frame.
The AI Disputes Monitor is J.S. Held's proprietary quarterly dashboard. The Q2 2026 edition was released from New York on July 15.