ROSS says it will ask Supreme Court to review Third Circuit's AI-training fair-use ruling

What happened: ROSS Intelligence, through its counsel Yar Chaikovsky of White & Case, announced it will seek Supreme Court review of the September 29, 2026 Third Circuit decision in Thomson Reuters Enterprise Centre GmbH v. ROSS Intelligence Inc., No. 25-2153. The panel (Judge Tamika Montgomery-Reeves) affirmed Judge Stephanos Bibas's February 2025 summary judgment that ROSS infringed 2,243 Westlaw headnotes — short summaries of legal holdings written by Thomson Reuters editors — by copying them to train its AI legal-search tool, and that the copying was not fair use. It is the first federal appellate decision to address fair use in the context of AI training.

Why it matters to creators: Supreme Court review could produce the first high-court guidance on AI training and fair use, shaping the licensing market for training data and the leverage of rightsholders in pending generative-AI cases. The Third Circuit opinion already gives rightsholders appellate authority that intermediate training copies can infringe even when protected text never appears in outputs, and that an emerging AI-training licensing market counts under fair-use factor four.

What is not yet known: the certiorari petition had not been filed as of the announcement — only the intent to seek review was stated. The appellate opinion remains under seal while the parties propose redactions. The panel also expressly distinguished generative AI in a footnote (discussing Bartz v. Anthropic, Kadrey v. Meta, and the DOJ's September 2026 statement of interest in the OpenAI multidistrict litigation), so the ruling does not settle the fair-use question for large language models.