Full Breakdown
Appeals Court Upholds Thomson Reuters Victory Over Ross Intelligence AI Training Dispute
By Drooid · · How we work
Court Ruling
The Philadelphia-based U.S. Court of Appeals for the Third Circuit rejected Ross Intelligence’s claim that its AI-driven legal search tool qualified for fair-use protection when it incorporated Thomson Reuters’ Westlaw headnotes. The panel’s opinion, filed as a sealed document, affirmed the lower court’s decision that Ross’s use was not “transformative” and therefore did not meet the fair-use standard.
Legal Background
Thomson Reuters filed a copyright lawsuit in 2020 alleging that Ross Intelligence copied thousands of Westlaw headnotes—concise summaries of legal points in court opinions—to train its competing AI search engine. After a federal judge in Delaware denied Ross’s fair-use defense last year, Ross shut down its platform in 2021, citing the cost of litigation. The present appeal marks the first time a U.S. appeals court has ruled on a copyright dispute involving AI training, though the case does not involve generative AI that creates new content.
Official Reactions
The spokesperson’s comment reflects the company’s broader strategy to safeguard its legal research assets.
Implications for AI Training
The decision reinforces copyright owners’ ability to control the use of their proprietary data in AI development, signaling that courts may scrutinize claims of transformation more narrowly. Tech firms that rely on large datasets of copyrighted material to train AI models could face heightened legal risk, potentially prompting more cautious data-acquisition practices or increased licensing negotiations. The ruling also adds to a growing wave of high-stakes copyright cases that could shape the regulatory landscape for AI across multiple industries.