The court found that Ross Intelligence had engaged in actual copying, but left it to a jury to decide the validity of Thomson Reuters' copyright and whether there was substantial similarity. The court also denied summary judgment on Thomson Reuters' direct, vicarious, and contributory infringement claims. The court's framing of Ross' fair use defense suggests that unauthorized use of third-party content to train an AI model could be a fair use where the model is not designed merely to replicate the original copyrighted content. The trial is tentatively scheduled for May 2024.
Key takeaways:
- The court case Thomson Reuters v. Ross Intelligence provides insight into how courts might address the issue of whether the use of copyrighted works to train artificial intelligence systems constitutes copyright infringement.
- The court denied summary judgment on Thomson Reuters’ direct, vicarious and contributory infringement claims, leaving many factual issues to be decided by a jury.
- The court established a legal framework that weighs heavily in favor of finding fair use if an AI system studies language patterns in copyrighted works to learn how to produce responses, and not to replicate the copyrighted content itself.
- Thomson Reuters also brought tortious interference claims against Ross, asserting that Ross induced a third party to breach the Westlaw licensing terms. The court denied summary judgment on these claims, allowing them to proceed to trial.