TL;DR
The Third Circuit has become the first U.S. appeals court to rule that using copyrighted material to train an AI system is not fair use, handing Thomson Reuters a landmark win over Ross Intelligence.
What happened
- The U.S. Court of Appeals for the Third Circuit upheld Thomson Reuters' lower-court victory against Ross Intelligence on October 1, 2026.
- Ross had used Westlaw's copyrighted headnotes (legal summaries) to train a competing AI-powered legal research tool.
- A three-judge panel rejected Ross's fair use defense, affirming that converting the summaries into AI training data did not radically transform them.
- The opinion was initially filed under seal before becoming the first U.S. appellate ruling directly addressing AI training and copyright fair use.
- The panel had signaled skepticism at oral argument that building a competing product from protected content could qualify as fair use, even as an intermediate step.
Why it matters
- This is the first appellate precedent in the U.S. establishing that AI training on copyrighted content can constitute infringement, not a transformative fair use.
- Every AI company ingesting third-party copyrighted data for model training now faces a materially stronger legal headwind in the Third Circuit and beyond.
- The ruling validates content owners (publishers, news organizations, legal data providers) who are demanding licensing fees or suing AI developers.
- Ross's core argument, that training data use is merely a functional intermediate step, was explicitly rejected, closing a widely hoped-for loophole.
- Thomson Reuters and Westlaw emerge with a reinforced competitive moat: rivals cannot cheaply replicate their curated legal corpus via AI training.
What to watch next
- Whether other circuits adopt or split from this reasoning, which could force a Supreme Court resolution of AI training copyright questions.
- How AI developers respond: expect accelerated moves toward licensed data deals or synthetic data strategies to avoid infringement exposure.
- Any legislative response in Congress, where AI and copyright reform bills have been circulating, now facing new urgency from binding appellate law.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.