AI in Action: Reimagining Metadata and Cataloging with Chatbots and OpenAI

Machine learning and Natural Language Processing (NLP) in Artificial Intelligence (AI) have significantly impacted libraries’ technical services practices, including content extraction, classification, and text generation. Chatbots and the OpenAI API are being utilized by libraries to streamline metadata work and enhance workflows on a global scale. This presentation will share the University of Central Florida’s experiences with using the OpenAI API to assign Faceted Application of Subject Terminology (FAST) headings and keywords to their digital collections in the Digital Commons-based Institutional Repository, as well as traditional collections such as theses and dissertations in Ex Libris’ Alma. In generating FAST headings, OCLC’s FAST Reconciliation Service was integrated into the AI dataflow and coding. This presentation will cover multiple rounds of testing that were conducted to refine the prompts and improve outcomes, as well as the development and application of evaluation rubrics. It will also address the ethical applications of AI and considerations of diversity and inclusion in AI content generation. The presentation will conclude with a discussion on next steps for applying AI in metadata and cataloging tasks, aiming to reimagine, refocus, and reset the paths for libraries to enhance resource discovery and user engagement for the future.