The U.S. Court of Appeals for the Third Circuit, based in Philadelphia, has affirmed a lower-court judgment in favor of Thomson Reuters in a copyright lawsuit against Ross Intelligence. The dispute centered on whether Ross used copyrighted materials to train its artificial intelligence–based legal search engine.
Thomson Reuters filed the suit in 2020. The complaint alleged that Ross copied several thousand “headnotes” — the legal summaries of court opinions — from the Westlaw system and used those summaries to train its own search tool. Ross Intelligence shut down its platform in 2021, citing the high costs of litigation among the reasons for the closure.
A district court concluded last year that Ross’s use of the headnotes was aimed at facilitating the development of a competing legal research product and therefore was not transformative. On that basis the district court found the use did not qualify as fair use. The appeals court has now affirmed that finding.
The appeals court’s full written opinion has not yet been released, so the detailed legal reasoning will be available when the opinion is published.
Why the case matters
-
This is the first U.S. appellate decision to address a copyright dispute connected to AI training, a ruling that may have broad implications for the technology sector because how courts interpret fair use affects how companies may reuse copyrighted works for training models.
-
The Thomson Reuters case does not concern generative AI that produces new expressive content. Instead, it involved a search engine that, after training, identified relevant court decisions in response to user queries.
Industry debate continues over whether AI systems transform copyrighted works or effectively replicate and compete with rights-holders’ products. The publication of the appeals court’s full opinion will be closely watched for guidance on those questions.
(Source: Reuters.)
