PHILADELPHIA, Pennsylvania — A federal appeals court has upheld a landmark copyright ruling against Ross Intelligence, rejecting the company’s argument that it made fair use of material from Thomson Reuters’ Westlaw platform to train a competing artificial-intelligence legal search engine.
The Third U.S. Circuit Court of Appeals issued the decision Tuesday in Thomson Reuters Enterprise Centre GmbH v. Ross Intelligence Inc., case 25-2153. Reuters reported that it is the first U.S. appellate ruling in a copyright case involving artificial-intelligence training.
Thomson Reuters accused Ross of copying thousands of Westlaw “headnotes,” which summarize points of law in court opinions, to develop its search system. A federal judge in Delaware previously ruled that Ross used the material to make development of a competing legal-research product easier and that the use was not transformative.
The decision is a significant signal for publishers, news organizations, music companies and other copyright owners pursuing artificial-intelligence developers over training data. Its reach should not be overstated, however: the Ross system was a legal search engine, not a generative-artificial-intelligence service that creates new text, images or music.
What remains unresolved
The appeals court’s reasoning remains under seal, Reuters reported. Until the opinion is made public, the precise legal analysis and its application to other artificial-intelligence training cases cannot be fully assessed. Ross had not publicly responded to the ruling at the time of publication.
Thomson Reuters said respecting copyright is essential both to innovation and to protecting the intellectual property used in professional artificial-intelligence products.