Encyclopedias Sue AI Over Training Data

Traditional reference publishers are beginning formal legal challenges against AI labs, claiming models were trained on massive troves of copyrighted articles and that generated outputs reproduce or falsely attribute their content. These suits combine copyright and trademark claims and seek injunctions as well as damages, signaling a coordinated industry response to generative AI's business and discovery impacts. — If successful, these cases could force changes to how models are trained, how companies license text, and how online search and traffic economics work — affecting consumers, publishers, and AI firms.