On February 12, a Delaware federal court rejected ROSS Intelligence Inc.'s fair use defense for using copyrighted material to train its artificial intelligence (AI) program in the bellwether AI copyright case, Thomson Reuters Enterprise Centre GmbH et al. v. ROSS Intelligence Inc. Intellectual Property Partner and head of Katten's AI Working Group Michael Justus was quoted by Law360 on the ruling, which is regarded as potentially influential in ongoing disputes over AI training as it relates to key issues in the fair use analysis such as market impact and transformative use.
The ruling held, under the fourth fair use factor, that there was a potential market to license data from Thomson Reuters' Westlaw platform for AI training, which weighed against fair use. Mike commented that this aspect could be influential for future cases, as it underscores the importance of potential licensing markets for training data. "There's been highly publicized deals where some rights holders have chosen to license materials for AI training," he said. "So this is an argument you're going to see every plaintiff make — that in their particular industry, whatever that may be, they or someone else is either already licensing data for AI training or has the potential to do so."
The decision also concluded that ROSS' use of Thomson's content was not transformative — the first fair use factor — since it served a similar commercial purpose as Westlaw (i.e., to develop a competing legal research tool). On this, Mike noted, "Because of the way factor four came out, which suggests that if there's a potential market for licensing training data that weighs in favor of plaintiffs, it may turn out that factor one becomes the bigger focus in the other generative AI cases because it won't be nearly as clear in some instances that the uses are the same."
"4 Takeaways From 1st Opinion On AI Training And Fair Use," *Law360, February 13, 2025
*Subscription may be required for article access.
Washington, DC
