US appeals court upholds Thomson Reuters win against Ross Intelligence on Oct 18, reinforcing AI training fair use precedent
The Third Circuit Court of Appeals upheld the February 2025 district court ruling against Ross Intelligence, holding that scraping copyrighted Thomson Reuters legal data to train AI is not fair use; the opinion remains sealed.
On October 18, the US Third Circuit Court of Appeals in Delaware upheld the February 2025 district court ruling by Judge Bibas against Ross Intelligence, holding that the company's unauthorized scraping of Thomson Reuters' Westlaw legal database to build a legal AI did not qualify as fair use. The appellate opinion remains sealed, but Reuters characterized the decision as a landmark ruling in the AI training copyright space.
Ross Intelligence argued its use of Westlaw data to train a legal research AI tool was transformative and thus protected by fair use. Neither court adopted the argument. The ruling is seen as a favorable precedent for ongoing copyright suits by writers' guilds and news publishers against OpenAI, Anthropic, Meta, and other AI companies.
The appellate affirmance came alongside compliance moves from AI companies responding to data-source scrutiny, including Meta's October 17 removal of language from its commercial terms that critics said could have allowed scraping of news sites. Legal experts expect the appellate ruling to influence pending copyright cases against OpenAI and Microsoft, Anthropic and multiple publishers, and others.