This summary is based on the Court of Appeals for the Third Circuit’s published decision in Thomson Reuters Enterprise Centre GmbH v. Ross Intelligence Inc, docket number 25-2153, filed September 30, 2026.

The court addressed an interlocutory appeal arising from the alleged copyright infringement by Ross Intelligence, Inc. (ROSS) of editorial materials created and owned by Thomson Reuters Enterprise Centre GmbH (Thomson Reuters), specifically Westlaw headnotes. Thomson Reuters accused ROSS of copying 2,243 Westlaw headnotes to train an artificial intelligence (AI) legal research platform that competes with Westlaw.

Westlaw headnotes are editorial summaries and annotations that explain points of law in judicial opinions, created by Thomson Reuters editors under precise guidelines to be clear, concise, and independent of the underlying opinion text. Thomson Reuters holds a registered copyright in Westlaw’s materials, including the headnotes, the Key Number System used to index legal topics, and textual synopses.

ROSS developed an AI legal search engine that answers plain-language legal questions by retrieving relevant excerpts from uncopyrighted judicial opinions. To improve its AI, ROSS enlisted a contractor, LegalEase Solutions, to prepare approximately 25,000 legal memoranda using Westlaw headnotes as frameworks for legal questions and identifying judicial opinion passages of varying relevance. ROSS then used these memoranda to train its AI system. The parties agreed that ROSS copied the headnotes in this process and that such copying was attributable to ROSS.

Thomson Reuters sued ROSS for copyright infringement and tortious interference with contract. The district court granted partial summary judgment in favor of Thomson Reuters, ruling that the headnotes possess sufficient originality for copyright protection and that ROSS’s copying was not fair use. The district court subsequently certified two questions for interlocutory review: whether the Westlaw headnotes and the Key Number System are original as a matter of law, and whether ROSS’s use qualifies as fair use.

The Third Circuit affirmed the district court’s ruling. Regarding originality, the court held that Westlaw headnotes meet the extremely low threshold of copyright originality, possessing a “creative spark” due to editorial choices about which legal points to include and how to express them clearly and accurately. The court rejected ROSS’s merger doctrine argument, explaining that many different ways exist to express points of law in headnotes. The court also found unpersuasive ROSS’s reliance on a Second Circuit case that did not consider headnotes unoriginal.

“Because these decisions show the requisite minimal ‘creative spark,’ the headnotes are original.”

On the fair use defense, the court analyzed the statutory four-factor test. First, it found ROSS’s use commercial and minimally transformative, as both Thomson Reuters and ROSS use the headnotes to serve similar purposes in legal research platforms. The court distinguished the case from precedent involving generative AI and transformative uses in other contexts, noting that ROSS’s AI does not generate original expressions and that ROSS aimed to compete directly with Westlaw.

Second, the court found that the nature of the copyrighted work favored fair use due to the factual nature of the headnotes and their public availability. Third, the court concluded that the amount copied (entire headnotes) was excessive relative to ROSS’s nontransformative purpose, especially since ROSS could have used uncopyrighted judicial opinions instead.

Fourth, the court found the market effect factor weighed against fair use, concluding that ROSS’s copying harmed the value of Thomson Reuters’s headnotes by appropriating their utility to attract legal research customers and by preempting derivative markets involving licensing headnotes as AI training data. The court also determined that alleged public benefits claimed by ROSS, such as increased access to law or AI advancement, were unsubstantiated or less important in this context.

Consequently, the Third Circuit held that Westlaw headnotes are copyrightable works and that ROSS’s use was not fair, affirming the grant of partial summary judgment in favor of Thomson Reuters.

Editorial: Potential Impact

This decision may have significant implications for both the legal and AI industries by clarifying that editorial legal materials, such as Westlaw headnotes, possess copyright protection even when used in AI training. The ruling signals that courts may closely scrutinize AI developers’ use of copyrighted third-party content, especially when the new use is commercial and not sufficiently transformative.

By affirming that copying entire headnotes for training purposes does not qualify as fair use absent a transformative purpose, the court may encourage AI companies to seek proper licensing arrangements when using proprietary legal editorial content. The recognition of derivative markets for licensing headnotes as AI training data also suggests an emerging commercial avenue that providers like Thomson Reuters may pursue, potentially affecting how AI tools for legal research are developed and marketed.

Given this decision is an interlocutory appeal from the Third Circuit, its scope is narrow and fact-specific. However, it underscores ongoing tensions between copyright law and AI innovation, which may prompt further legislative or judicial attention as AI technologies continue to evolve and expand in the legal field.

Source: Read the full decision on CourtListener

Court: Court of Appeals for the Third Circuit | Docket: 25-2153 | Date filed: 2026-09-30

This article was generated automatically from a public court record and is provided for informational purposes only. It is not legal advice. Refer to the linked source document for the authoritative text of the decision.