A US appeals court upheld a significant copyright ruling in favor of Thomson Reuters, rejecting a challenge by legal research firm ROSS Intelligence regarding the use of Westlaw’s content to train artificial intelligence systems. The court dismissed ROSS’s defense that its use constituted fair use, strengthening the protection of copyrighted legal databases from unauthorized AI training.
The dispute centers on whether AI developers can incorporate proprietary databases into their training data without explicit permission. Thomson Reuters argued that ROSS’s AI training on Westlaw’s material infringed on their copyrights, while ROSS claimed their use was transformative and covered under fair use. The appeals court disagreed, highlighting that training AI on protected content without consent does not meet fair-use criteria, particularly when the content plays a commercial role.
This ruling clarifies a contentious area of copyright law involving AI development, especially as large datasets are critical for training machine-learning models. The court’s decision signals that companies providing proprietary data can enforce copyright protections more strictly against unauthorized AI data scraping.
The case has broader implications for the AI industry, particularly for companies relying on third-party datasets to develop search algorithms or other AI-driven services. It establishes legal precedent emphasizing that fair use defenses may not apply when copyrighted content feeds directly into commercially oriented AI training models.
Thomson Reuters’ Westlaw is a comprehensive legal research platform, and its protection under copyright law is crucial in preserving the value of licensed information services. This decision may influence how AI developers approach data sourcing to avoid infringement risks.