Peter Leonard
Associate University Librarian for Research Data Services
Stanford University
What can academic libraries offer researchers in a world where everyone uses ChatGPT? The talk explores artificial intelligence (AI) modeling and inference as a service, articulated through hardware (local GPUs), software (open-source models and software), and professional expertise (consulting, design, and implementation). It will detail Stanford University Libraries’ new unit devoted to this task, exploring job descriptions, service workflows, and hardware and software infrastructure. It addresses both faculty and graduate student research projects and engagement with large-scale digitized collections of text, image, and audio/video. It also considers adjacencies with subject specialists, library IT, and research computing groups on campus.