US Landmark AI Copyright Case Could Reshape Intellectual Property Law

AI companies may face stronger copyright challenges following a pivotal legal ruling.

A U.S. federal judge recently ruled in favor of Thomson Reuters in its lawsuit against Ross Intelligence, finding that Ross improperly used Reuters' proprietary legal content to train its AI research platform.

This precedent-setting decision could influence dozens of pending AI copyright cases nationwide, though legal experts caution it doesn't guarantee wins for all plaintiffs.

The significance of legal headnotes

The case centered on Ross' use of Westlaw's copyrighted legal headnotes—concise case summaries—to train its competing AI research tool.

Ross claimed its usage constituted transformative fair use, arguing it repurposed the content for different legal applications. Judge Stephanos Bibas rejected this defense in his summary judgment.

In his ruling, Bibas determined Ross essentially repackaged Westlaw's proprietary content without adding substantive analysis or commentary, directly competing with Reuters' core legal research business.

The judge further noted Ross' commercial motivations undermined its fair use claim, as the company sought to profit from Westlaw's intellectual property without sufficiently transforming it.

Join us at TechCrunch Sessions: AI

Reserve your spot at our premier AI conference featuring leaders from OpenAI, Anthropic and Cohere. Limited-time offer: Full-day access including expert sessions and networking for just $292.

Exhibit at TechCrunch Sessions: AI

Showcase your innovations to 1,200+ industry leaders at TC Sessions: AI. Limited exhibitor tables available through May 9.

Syracuse University IP law professor Shubha Ghosh hailed the decision as a significant win for Thomson Reuters.

"While the trial continues, Thomson Reuters secured an important preliminary victory," Ghosh noted. "The judge's rejection of Ross' key defenses establishes favorable momentum for the plaintiff."

Potential limitations

Though the ruling has been cited in other AI copyright cases, legal experts debate its broader applicability.

Judge Bibas explicitly distinguished Ross' case from generative AI lawsuits, noting Ross' system merely reproduced existing content rather than creating new material.

Generative AI systems—like those developed by OpenAI and Midjourney—analyze vast public datasets to produce original media, raising different copyright questions.

AI developers typically argue their data usage qualifies as fair use since models transform training materials into novel outputs. However, copyright holders counter that AI-generated content often closely replicates protected works.

Hall Estill attorney Randy McCarthy said Bibas' focus on commercial impact could strengthen rights-holders' positions, but cautioned the ruling remains narrowly focused.

"This establishes that using copyrighted material for AI training doesn't automatically constitute fair use," McCarthy explained. "But we'll need more rulings to fully clarify the legal landscape."

Knobbe Martens litigator Mark Lezama suggested the precedent could extend to generative AI disputes, particularly when AI outputs compete with original content sources.

While the decision offers modest encouragement for copyright holders fighting AI companies, legal experts stress the broader battle remains far from settled.