US Appeals Court Upholds Thomson Reuters' Copyright Win in AI Training Case

A U.S. appeals court on Tuesday upheld a landmark copyright ruling in favor of information services company Thomson Reuters, affirming that a rival firm, Ross Intelligence, illegally used copyrighted material to train its artificial intelligence-powered legal search engine. The decision from the Philadelphia-based 3rd U.S. Circuit Court of Appeals marks the first time a federal appellate court has ruled on a copyright dispute involving AI training, a pivotal issue for the rapidly growing technology sector.

The Core of the Dispute

The case stems from a 2020 lawsuit in which Thomson Reuters accused Ross Intelligence of copying thousands of its proprietary "headnotes"—short summaries of legal points from court opinions—from its Westlaw platform. Thomson Reuters alleged that Ross used this content to train its own competing AI legal-research tool. Ross Intelligence, which ceased operations in 2021 citing the cost of the litigation, argued its actions constituted fair use.

A Delaware federal judge rejected Ross's fair use defense last year, a decision the appeals court has now affirmed. The lower court judge stated that Ross's use was not transformative because it "took the headnotes to make it easier to develop a competing legal research tool." The reasoning behind the 3rd Circuit's new decision is currently sealed, according to reports.

Implications for AI and Copyright Law

Ad

This ruling arrives amid a wave of high-stakes lawsuits filed by authors, artists, and news organizations against technology companies over the use of their work to train AI systems. While the Thomson Reuters case does not involve generative AI—which creates new content—it directly addresses the critical legal defense of fair use that underpins many AI business models.

Key takeaways from the decision include:

• Strengthened Copyright Holder Position: The ruling reinforces the argument that using copyrighted data to train a directly competing product may not qualify as transformative fair use.
• A Key Legal Precedent: As the first U.S. appeals court decision on AI training, it will be closely scrutinized and likely cited in other ongoing copyright disputes, even those involving generative AI.
• Increased Scrutiny on Training Data: The verdict may prompt AI developers to be more cautious about the sources of their training data and could accelerate the market for licensed data.