Academic libraries are increasingly exploring how AI can enhance patron services while maintaining our commitment to privacy, accessibility, and information literacy. Here at UT Austin, we’re excited to share our ongoing research on a Library Assistant chatbot designed to support patrons with everything from questions about book lending policies and on-campus printing to research queries.
The chatbot project began in Fall 2023 as an MSIS capstone by Hannah Moutran, under the guidance of Aaron Choate, Director of Research & Strategy at UT Libraries. What started as a capstone has evolved into an ongoing research initiative that’s helping us understand how AI can thoughtfully support both library staff and patrons. The project was recently featured in the Coalition for Networked Information (CNI) August 2024 Pre-Recorded Project Briefing Series, where Hannah shared insights about the development process and early findings.
The team began by analyzing Fall 2022 chat logs to understand actual patron needs – revealing that 51% of interactions involved research assistance, 34% were about general library services, and 13% focused on accounts and technical support. Building on these insights, we conducted interviews with librarians who staff the chat service to understand the nuances of patron interactions and identify opportunities where AI could provide meaningful support.
The resulting prototype features several conversation flows, including a General Library Flow for questions about physical spaces and services and a Research Assistance Flow that helps users to brainstorm research topics and begin their search for online resources. Each aspect of the design draws from both our own research and Ethical AI guidelines, incorporating principles from the ALA Code of Ethics and Microsoft’s HAX framework.
This is very much a work in progress, and we’re continuing to refine the chatbot based on ongoing testing and feedback. If you’d like to try it out yourself, you’ll find a demo link below – and know that every interaction helps us better understand how to thoughtfully integrate AI into library services.
Access the prototype: UT Library Assistant Chatbot Demo
Questions or thoughts on the topic?
Get in touch!