Delaware Court Rules AI Training is not Fair Use

A major unresolved legal issue related to AI is whether using copyrighted works for training AI models violates copyright law. Many developers caught in copyright infringement suits argue the defense of “fair use.” However, a Delaware Court in the high profile Thomson Reuters v. ROSS Intelligence case recently sided with plaintiffs and issued an opinion declining to extend the “fair use” defense to AI training. A recent Debevoise & Plimpton memo breaks down the decision stating that:

“The most noteworthy aspect of the court’s ruling pertained to fair use, which has been at the forefront of many copyright cases involving generative AI. Reversing its earlier position that issues of fact precluded summary judgment on fair use for either party, the court held that ROSS’s activities were not fair use as a matter of law.”

While the ruling is an important early decision on AI and fair use, it isn’t dispositive of all AI related copyright claims. In this case, Ross Intelligence used Thomson Reuters headnotes to train a legal search engine powered by AI without acquiring the rights. As the memo points out the facts of this case are notably different from other AI copyright cases. The Ross Intelligence system was a search engine powered by AI, not a generative AI tool. Additionally, Ross Intelligence was using their product to directly compete against Thomson Reuters. Other pending cases may have significantly different facts that alter the outcome of the four prong fair use analysis. While the impact of Thomson Reuters v. Ross Intelligence shouldn’t be overstated, it does provide a first step towards legal certainty regarding AI and copyright law.