Artificial Intelligence, Authentic Reactions: Tensions Between Digital Libraries and Generative AI

abstract: Artificial intelligence (AI), large language models, and their associated crawlers are the newest users for university libraries, special collections, and institutional repositories. Generative AIs want data from library collections and repositories for training their models. In some cases, they ask for it through the robots.txt protocol. In other cases, they cause accessibility issues for repositories and create copyright concerns. This article investigates these issues and offers potential solutions, including donor outreach and communication, working within the system, and enacting restrictions. Ultimately, we conclude that the development of sustainable, ethical, and effective strategies for dealing with AI cannot rest on the shoulders of individuals. Digital librarians, archivists, professional societies, and working groups must have conversations that work through these tensions between digital collections and AI tools while still advocating for our digital creators and meeting our patrons’ research needs.