Hannah Moutran
Research Assistant
University of Texas at Austin
This presentation shares the prototype development of an academic library assistant chatbot for the University of Texas at Austin Libraries, intended as an after-hours extension of the human-staffed Ask a Librarian service that aligns with the values and needs of the library system. Initial work included analyzing historical chat logs, interviewing five librarians who staff the Ask a Librarian service, evaluating different services for prototype design, and deciding on Voiceflow. After creating the first iteration of the chatbot, it was evaluated through the lens of two functional and ethical rubrics: Microsoft’s Human-AI Interaction (HAX) Guidelines and the American Library Association’s Code of Ethics. The chatbot design blends traditional conversation design with large language model calls made in specific steps. It is tailored to the library’s unique requirements; for example, it is designed not to answer questions directly but to guide users to resources and build search links with keyword parameters to create a research starting point. This method minimizes the need for updates, provides reliable information, and fulfills the library’s role as a gateway to information, fostering a user-friendly educational experience for students and faculty alike. The briefing provides insights that other institutions may use to guide the implementation of a customized library chatbot that provides high-quality research assistance to community members at all hours while also actively working to ensure the privacy of the community’s research pursuits.
The project was developed as Hannah Moutran’s capstone with her supervisor, Aaron Choate.
A blog post on the project: https://liblablexicon.com/2024/04/22/enhancing-library-services-with-conversational-ai/
Test the Library Assistant Chatbot: https://creator.voiceflow.com/prototype/661004fd5123b84a805c0724