Institutional Repositories as Publishers–Toward a Responsive Generative AI Policy

Conversation around the impact of generative AI and institutional repositories (IR) can focus on the implications of AI companies harvesting IR publications for training purposes. But when we think about IRs as publishers, we confront another set of questions. In response to the advent of tools like Chat GPT, scholarly publishers have released policies of varying permissiveness, circumscribing the use of generative AI tools in the writing, editing, and overall production of manuscripts. IRs inhabit a unique space relative to these publishers. In many cases, they provide a platform for previously-published articles. Such articles are now often the subject of generative AI policies upstream. In other cases, IRs publish original manuscripts, theses, dissertations, journals, and a host of other materials, some of which might now have employed generative AI tools in their creation. How might IR managers respond to the increasing use of generative AI in research publication with appropriate policies for use and disclosure, taking into account the unique position of IRs within the scholarly communications landscape? These are questions we are thinking about at Dartmouth and which I will explore in this talk.