Investigating Librarian Perspectives on Generative AI

This study investigates librarians' perspectives on generative artificial intelligence (AI) in libraries. Librarians (n = 52) from 21 different institutions attended a workshop to discuss generative AI and assess the ethics of various applications of large language models (LLMs) in research. Through the analysis of qualitative workshop data, we identified 24 overarching themes — 15 concerns related to embracing AI, 4 about not embracing it, and 5 possible interventions for libraries. Additionally, assessments revealed a nuanced view of the ethical considerations associated with LLMs in research. High ethical ratings were given to use cases involving language feedback and learning support. Lower ratings were associated with generative AI imitating professional roles and generating text based on user input. Furthermore, we found mixed opinions on several use cases among the participants, demonstrating the difficulty of establishing clear guidelines. This paper presents the preliminary findings from a research and development project, offering insights into the perspectives of librarians on generative AI. It underscores the need for a balanced, informed approach to integrating generative AI in libraries.