Third Circuit ruling in Thomson Reuters v. ROSS recognizes AI training data market

Baker Botts reports that on September 29, 2026 the U.S. Court of Appeals for the Third Circuit issued a decision in Thomson Reuters v. ROSS that it characterizes as apparently the first federal appellate decision addressing whether training an AI system on copyrighted material qualifies as fair use; the firm's headline takeaway is that the court recognized a licensing market for AI training data.

For counsel advising AI developers, data licensors, publishers, and law-firm content owners, the significance is that a federal appellate court has treated AI training data as a legally cognizable licensing market rather than treating the use as categorically exempt — which can influence negotiation posture, licensing program design, and the shape of copyright claims and defenses in pending disputes. The published summary, however, does not set out the panel, the full reasoning, or which fair-use factors carried the outcome, and the firm itself hedges that this is the first such appellate decision. Watch for further coverage of the opinion, petitions for rehearing or certiorari, and parallel litigation in other circuits before treating the ruling as settled nationwide.