Publications and Research
Document Type
Presentation
Publication Date
4-3-2025
Abstract
In Spring 2023, the Leonard Lief Library at Lehman College (CUNY) conducted an assessment of its library chatbot which uses OpenAI’s ChatGPT API with the goal of improving its functionality and performance. This presentation highlights the assessment’s findings, including types of questions asked, and areas where the chatbot succeeds and where it struggles. Additionally, the presenters will detail the creation of a rubric to measure the accuracy and completeness of the chatbot’s answers for users. Through this presentation, the authors will showcase best practices for libraries looking to implement or experiment with chatbots at a time when use of artificial intelligence in libraries is increasing. Objective 1: Define assessment in terms of accuracy and completeness in measuring a chatbot’s answers to users Objective 2: Understand ChatGPT’s strengths when used in a vendor managed chatbot model Objective 3: Critically consider a chatbot*s place in as a library website enhancement
