The United States Court of Appeals for the Third Circuit has upheld Thomson Reuters’ copyright victory against ROSS Intelligence in a major dispute involving the use of Westlaw legal materials to develop an artificial-intelligence-powered legal research platform.
The appellate decision leaves intact a lower-court ruling that ROSS’s use of copyrighted Westlaw headnotes was not protected by the fair-use doctrine.
The litigation has attracted particular attention because it is among the earliest appellate decisions directly confronting copyright and AI training.
How the dispute began
Thomson Reuters sued ROSS Intelligence after alleging that the technology company obtained and used thousands of Westlaw headnotes — editorial summaries of judicial decisions — in developing a competing legal research service.
ROSS disputed the allegations and argued that its use was transformative and therefore protected by the doctrine of fair use.
The lower court ruled in favour of Thomson Reuters on key copyright issues.
The Third Circuit has now upheld that result.
Why the decision matters
The dispute is particularly significant because it concerns the use of copyrighted professional legal content to build an AI-enabled product operating within substantially the same commercial market.
That distinguishes it from many current lawsuits involving large generative AI systems trained on vast collections of books, journalism, artwork and other material.
The decision should nevertheless not be read as establishing that all AI training involving copyrighted material is automatically unlawful.
Fair-use analysis under United States copyright law remains highly fact-specific.
Courts typically consider factors including the purpose and character of the use, the nature of the copyrighted work, the amount used and the effect of the new use on the potential market for the original material.
Legal research industry watches closely
The commercial relationship between Thomson Reuters and ROSS was central to the dispute because ROSS was developing a legal research platform capable of competing with Westlaw.
That market effect distinguishes the litigation from cases where copyrighted works are used for substantially different purposes.
The judgment is therefore expected to be closely studied by AI developers, publishers, lawyers, copyright owners and legal-technology companies.
For the legal profession, the case carries particular significance because the protected works were not conventional literary or entertainment products but editorial material specifically produced for professional legal research.
As artificial intelligence becomes increasingly integrated into legal research and professional services, the judgment is likely to form part of the developing international jurisprudence on where innovation ends and copyright infringement begins.
Related Stories
READ MORE: AI Copyright Fight Puts Fair Use Before US Court Nigerian Law Updates
READ MORE: Navigating Nigerian Copyright Law in the Age of Artificial Intelligence: Emerging Challenges and Proposed Solutions Nigerian Law Updates
Source: Reuters
