Court Rules Against Ross Intelligence in Copyright Case
The ruling restricts AI developers from using proprietary legal headnotes to train models without licensing.
Updated on Oct. 26, 2026 in Artificial Intelligence
Live Poll
Should artificial intelligence companies be required to obtain permission before using copyrighted content for model training?
On September 29, 2026, the US Court of Appeals for the Third Circuit ruled that Westlaw headnotes are protected by copyright and rejected a fair use defense by Ross Intelligence. The decision follows Ross Intelligence's 2021 shutdown of its search engine due to the legal costs of the dispute.
Why it matters
The ruling clarifies that legal research platforms cannot freely ingest copyrighted headnotes to train artificial intelligence models. This decision sets a restrictive precedent for developers building specialized AI tools on proprietary legal datasets.
The court analyzed the case based on the 4 factors of fair use to determine if using proprietary legal summaries for machine learning constitutes infringement. The bench affirmed that Westlaw headnotes maintain copyright protection despite their use as training data.
The players
Ross Intelligence
A legal technology company that developed an AI-powered search engine before ceasing operations in 2021.
Thomson Reuters
The parent company of Westlaw, a primary provider of legal research software and proprietary legal databases.
The details
Ross Intelligence attempted to use Westlaw headnotes as training data to develop an artificial intelligence program capable of querying case law. The Third Circuit rejected the company's fair use defense, which claimed that repurposing these legal summaries was transformative. By ruling that these specific research aids are protected by copyright, the court prevented competing platforms from using them as foundational material for legal AI.
Timeline
2021: Ross Intelligence shuts down its search engine due to litigation costs.
September 29, 2026: The US Court of Appeals for the Third Circuit issues its ruling.
The Tech Race
This ruling significantly narrows the scope of the fair use defense as it pertains to the development of legal artificial intelligence. It confirms that established incumbents can successfully block competitors from ingesting their proprietary databases for model training.
Legal professionals and researchers can expect less competition from AI platforms attempting to scrape proprietary database content for training. Developers looking to build legal-focused tools must now secure licensing agreements rather than relying on the fair use of existing legal headnotes.
The takeaway
The court has effectively set a high barrier for AI startups aiming to build competing legal research platforms using existing, proprietary annotations. Watch for how other copyright holders in the legal and academic sectors utilize this ruling to defend their training datasets from similar AI ingestion.
Further reading
For broader trends in machine learning regulation, visit the Artificial Intelligence section.
Source note: This article includes information reported by Music Connection Magazine.
Live Poll
Should artificial intelligence companies be required to obtain permission before using copyrighted content for model training?
