Court Rules Against Ross Intelligence in Westlaw Copyright Case

On February 11, 2025, Judge Stephanos Bibas issued a key ruling in a legal battle between Thompson Reuters (Westlaw) and Ross Intelligence. Thompson Reuters Enterprise Centre GmbH et al v. ROSS Intelligence Inc., No. 1:20-cv-613.  The case centered on whether Ross illegally used copyrighted legal summaries from Westlaw to train its AI search tool. The court ruled in favor of Thomson Reuters, finding that Westlaw’s Key Number System and over 2,200 headnotes were original enough to be protected by copyright law.

The judge rejected Ross Intelligence’s argument that its use of Westlaw’s headnotes was fair use. Two major factors worked against Ross: its AI tool was a commercial product, and its use of Westlaw content harmed Thomson Reuters’ business. Because Ross’s AI tool directly competed with Westlaw’s legal research services, it posed a threat to its market.

The court determined that Westlaw’s Key Number System and headnotes met the legal requirements for copyright protection. Ross had copied 2,243 of 2,830 Westlaw headnotes, and its training documents closely resembled the original content. The court rejected Ross’s defenses such as fair use, innocent infringement, and copyright misuse.

A crucial part of the ruling was Judge Bibas’s distinction between Westlaw’s headnotes and software. The court found that Westlaw’s headnotes were not software but rather literary works with a functional purpose. This distinction proved pivotal in rejecting Ross’s fair use claim.

The ruling suggests that using copyrighted content for AI training — even when not directly replicating it — may still constitute infringement. Even indirect use of copyrighted material can lead to legal trouble.

As AI technology evolves, future court cases may further define the limits of using copyrighted content for AI training. AI developers must tread carefully when using proprietary data to train their models.