Introduction
The Third Circuit Court of Appeals is poised to review a pivotal case involving ROSS Intelligence Inc and Thomson Reuters, focusing on the intersection of copyright law and artificial intelligence. This case [1] [5], ROSS Intelligence v Thomson Reuters, addresses the contentious issue of whether the use of copyrighted materials for AI training constitutes fair use, with significant implications for the legal and technological sectors.
Description
The Third Circuit is set to review an appeal by ROSS Intelligence Inc concerning a district court ruling that found the company liable for infringing on content from Thomson Reuters’ Westlaw legal research database. The district court determined that ROSS’s use of Westlaw’s headnotes to train its AI-powered legal research tool does not qualify as fair use under copyright law. This case [1] [5], ROSS Intelligence v Thomson Reuters, is significant as it represents the first summary judgment ruling to reject a fair use defense in the context of AI training and copyright issues [2]. The appellate court has granted ROSS’s interlocutory appeal regarding the rejection of fair use and the copyrightability of Westlaw’s headnotes and the Key Number System. Judge Cindy K Chung [2], along with Judges Peter J Phipps and D Brooks Smith [2], signed the order for the appeal [2], recognizing the importance of these matters and agreeing that they merit review.
Previously, the district court ruled that ROSS infringed copyright by utilizing Westlaw’s headnotes to develop a competing legal research tool [3], primarily rejecting ROSS’s fair use defense due to the competitive nature of the resulting AI product [3]. Critics have raised concerns about the court’s reasoning, particularly regarding the copyright protection of Westlaw’s headnotes [5], which largely consist of paraphrased judicial texts [5]. The court contends that these elements are original and thus copyrightable; however [5], this interpretation conflicts with established copyright principles that necessitate a degree of human creativity, which is notably absent in the headnotes.
Moreover, the court’s application of the merger doctrine—asserting that expressions of facts are not copyrightable when there are limited ways to express them—has been called into question. The claim that there are numerous ways to articulate legal points overlooks the reality that precise legal language often restricts expression options [5]. This misinterpretation could impose unnecessary limitations on the discussion and dissemination of legal information [5].
The court’s assertion regarding the originality of the selection and arrangement of the Key Number System is also under scrutiny, as it relies on automated processes rather than creative judgment [5]. The compilation of headnotes [5], which consists of uncopyrightable materials [5], does not warrant separate copyright protection [5].
In evaluating ROSS’s use of Westlaw’s materials [5], the court’s stance on intermediate copying is problematic [5]. It suggests that the existence of non-public intermediate copies constitutes copyright infringement [5], despite established case law indicating that such copies [5], particularly when not intended for public distribution [5], favor fair use [2] [3] [4] [5]. Previous rulings have acknowledged the importance of intermediate copying in the creative process [5].
Furthermore, the fourth fair use factor [5], concerning market substitution [5], has been critiqued for its circular reasoning [5]. The court’s view that ROSS’s use of Westlaw’s data creates a market for AI training data that should be controlled by Westlaw is inconsistent with the principle that fair use should not be determined by speculative market impacts [5]. Overall, the ruling raises significant questions about the application of copyright law in the context of AI and legal technology [5], and the Third Circuit’s findings could potentially reshape the legal landscape surrounding these issues. A determination that the headnotes lack protectability may eliminate the need to evaluate the fair use defense [3], while a ruling on fair use could provide essential clarity regarding its application in AI training contexts [3]. This case reinforces previous Supreme Court rulings related to copyright and fair use [1], particularly in the context of AI and legal technology [1] [5].
Conclusion
The outcome of the Third Circuit’s review could have far-reaching consequences for the future of AI development and the legal industry. A decision in favor of ROSS Intelligence may pave the way for broader applications of fair use in AI training, while a ruling upholding the district court’s decision could reinforce the protection of copyrighted materials in the digital age. This case underscores the ongoing tension between innovation and intellectual property rights, highlighting the need for clear legal frameworks in the rapidly evolving field of artificial intelligence.
References
[1] https://www.worldipreview.com/copyright/fed-circ-grants-first-fair-use-ai-and-copyright-appeal-in-thomson-v-ross
[2] https://news.bloomberglaw.com/in-house-counsel/prewrite-do-not-publish-34
[3] https://perspectives.fbm.com/post/102kgk5/westlaw-headnotes-case-heads-to-appeals-court-in-closely-watched-ai-copyright-bat
[4] https://chatgptiseatingtheworld.com/2025/06/17/third-circuit-grants-ross-intelligence-interlocutory-appeal-of-fair-use-copyrightability-of-headnotes-ruling-by-judge-bibas/
[5] https://www.authorsalliance.org/2025/02/13/thomson-reuters-v-ross-the-first-ai-fair-use-ruling-fails-to-persuade/
