Encountering a Librarian-Like AI: Understanding Users' First Encounters in a Public Library

Abstract

Public libraries are increasingly introducing conversational artificial intelligence (AI) services, yet little is known about how users experience these services during their first encounters. While previous research has focused primarily on system performance and technical capabilities, this study examines how users recognize, interpret, and make sense of an emerging AI-mediated library service. This exploratory qualitative study investigated Xiaoshu, a librarian-like AI interface deployed in the National Library of Public Information in Taiwan. Eleven adult library users completed scenario-based information-seeking tasks using Xiaoshu for the first time, followed by semi-structured interviews. Observation notes, screen recordings, and interview data were analyzed using reflexive thematic analysis.

The findings show that user experience began before interaction with the AI itself. Participants first evaluated the service, formed expectations based on its appearance and context, adapted their communication strategies during use, and ultimately positioned Xiaoshu alongside librarians and existing library services. Rather than simply interacting with AI, participants were making sense of a new library service.

The findings suggest that implementing conversational AI is as much a service design challenge as a technological one. Understanding users' first encounters with AI services, including their human–AI interactions, can help libraries design, introduce, and...