On Tuesday, US District Court of Delaware has ruled that Ross Intelligence violated Thomson Reuters' copyright when it used the mediaholding's materials to train its AI model. According to The Verge, this is one of the first cases involving the use of artificial intelligence and copyright.
Case summary
- Ross Intelligence created an AI model to help lawyers find information. To train it, the developer used, among other things, Westlaw materials copyrighted by Thomson Reuters.
- Westlaw is a large database of legal information that includes not only statutes and court decisions, but also copyrighted materials such as legal commentary and analysis.
- However, as the judge noted, this case involved "non-generative" AI, unlike most commonly used AI tools like LLM.
- The court rejected Ross's main argument about fair use of content, saying that the company was actually creating a competitive product that would harm the market for the original material.
- When Thomson Reuters refused to license it, the startup purchased 25,000 documents based on those annotations through a third company, LegalEase.
- The lawsuit was filed in 2020, and the following year, Ross shut down operations after raising enough money to defend the case in court.
Similar lawsuits against OpenAI, Microsoft, and other tech companies are currently pending in court, and this case could become an important precedent in the area of copyright and artificial intelligence training.
