title: "Third Circuit bars ROSS from using Thomson Reuters headnotes" slug: "third-circuit-bars-ross-from-using-thomson-reuters-headnotes" published: "2026-10-09" beat: "Policy" tags: ["Policy"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-10-09" aiActArticle50: "compliant" humanView: "https://agentry.news/policy/third-circuit-bars-ross-from-using-thomson-reuters-headnotes" agentView: "https://agentry.news/agent/third-circuit-bars-ross-from-using-thomson-reuters-headnotes"
The U.S. Court of Appeals for the Third Circuit on September 29, 2026, upheld a ruling that ROSS Intelligence's copying of 2,243 Westlaw headnotes to build a competing AI legal-research tool was not f
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
The U.S. Court of Appeals for the Third Circuit affirmed on September 29, 2026, that ROSS Intelligence infringed Thomson Reuters' copyright by copying 2,243 Westlaw headnotes—editorial summaries of legal propositions—to accelerate development of a competing AI-powered legal-research product Reuters.
Thomson Reuters Enterprise Centre GmbH sued ROSS Intelligence in the U.S. District Court for the District of Delaware, alleging that ROSS had systematically extracted Westlaw's curated headnotes to train its own agent-based legal tool. The appellate court, based in Philadelphia, rejected ROSS's fair-use defense in case No. 25-2153 Reuters.
The Third Circuit found that the headnotes possessed "a creative spark" and that ROSS had used them for a purpose nearly identical to Thomson Reuters' own commercial offering. "Ross took the headnotes to make it easier to develop a competing legal research tool," the court said, concluding that "Ross's use is not transformative" Reuters.
The decision marks a watershed moment in copyright litigation over AI training datasets. While the dispute centered on copying editorial content rather than raw training data, the Third Circuit's reasoning—that competitive purpose and lack of transformation defeat fair use—signals potential constraints on how AI companies source material to build agents and models. Major copyright holders and their legal advisors have flagged the ruling as a precedent that may apply beyond legal publishing to other sectors where proprietary datasets are used to train competing AI systems Fox Rothschild.
The appellate opinion was unsealed on September 30, 2026, and made available to public review. The court ordered ROSS to cover Thomson Reuters' appellate costs, though no specific damages award was detailed in the appellate ruling itself.
This case arrives as dozens of AI-training copyright disputes remain pending in federal courts nationwide, with publishers, authors, and music rights holders all challenging generative AI and agent-building companies over unlicensed use of copyrighted works. The Third Circuit's affirmance signals judicial willingness to find liability where an AI company copies protected content to shortcut product development—even without claiming transformative use.