The U.S. Court of Appeals for the Third Circuit ruled that Ross Intelligence violated Thomson Reuters’ copyright by using Westlaw’s annotations to train its own legal information search platform to compete with Westlaw.
Image credit: Tingey Injury / unsplash.com
Ross developed an AI-based legal search platform that allows attorneys to ask questions and retrieve relevant portions of court decisions. The company, along with third-party contractor LegalEase Solutions, prepared educational materials that linked legal issues to snippets of court rulings.
These materials include Westlaw Notes – concise summaries prepared by the platform’s editors that highlight the legal aspects of court decisions. Their task is to help lawyers quickly find passages in the judgment text that are relevant to specific legal issues. Ross argued that the annotations were too closely related to the original court decision to be protected by copyright, and insisted that they were used in good faith — to train artificial intelligence, not to be republished.
The court rejected both arguments and held that Westlaw’s notes met the minimum threshold of originality required for copyright protection: The editor decides what legal issues each note covers, what facts and arguments it contains, and how to accurately and succinctly present the point.
An important aspect is the status of Westlaw materials in the Ross system – they are not used for studying legal texts or internal research, but for training products that can answer questions within the framework of legal retrieval, which is also one of the main functions of Westlaw itself. Ross didn’t show the annotations to users, but that didn’t help—the court ruled that training artificial intelligence was only an intermediate stage in creating a competing commercial service.
Such uses of information are classified as “Minimal change at best”that is, Ross took a different technical approach but pursued similar business goals. The company leverages Westlaw’s editorial expertise to enhance the system and target the same market for legal information retrieval services.
An important aspect of the fair use doctrine is market harm, and the court held that Ross’s actions may affect the value of Westlaw’s services in this area and create obstacles to the potential market for licensed annotations as artificial intelligence training materials.
The outcome of the case highlights the importance of the source and purpose of materials used to train artificial intelligence – a developer can open public documents, but that does not give him the right to use the publisher’s annotations, classification tags, ratings or other editorial work based on those documents.
If you find an error, select it with your mouse and press CTRL+ENTER.
