
The first federal appeals ruling on AI training and fair use backs Thomson Reuters, but sets generative AI apart.
The 3rd U.S. Circuit Court of Appeals has ruled for Thomson Reuters in its copyright suit against ROSS Intelligence, affirming Judge Stephanos Bibas's February 2025 grant of partial summary judgment on 2,243 Westlaw headnotes, LawNext reported. The panel found the headnotes protected by copyright and rejected the argument that training ROSS's AI legal research tool on them was fair use.
The opinion, issued under seal on September 29 and later released, is the first by a federal appeals court on whether training an AI system on copyrighted material is fair use. Writing for the panel, Judge Tamika R. Montgomery-Reeves called it no more than an ordinary copyright case. The court stressed that ROSS's AI was not generative and only returned passages from existing opinions, distinguishing Bartz v. Anthropic PBC and In re OpenAI, Inc. Copyright Infringement Litigation.
According to the court, ROSS used the headnotes to build a commercial substitute for Westlaw, could have worked from the judicial opinions themselves, and by copying without a licence took from Thomson Reuters the chance to enter a rapidly developing market for licensing headnotes as AI training data. Judges L. Felipe Restrepo and Emil J. Bove sat on the panel.
Thomson Reuters sued in Delaware federal court in 2020; ROSS shut down its platform in 2021, citing litigation costs. Yar Chaikovsky, a White & Case partner, said in a statement on behalf of ROSS that it disagreed with the decision and intends to seek Supreme Court review. A Thomson Reuters spokesperson told Reuters the company was pleased with the ruling.
Published under the editorial responsibility of Kelly Moon, Editor-in-Chief



