Abstract

This presentation focuses on Gaby, an experimental AI chatbot developed to improve library information services at Concordia University. The project explores various techniques, including prompt engineering, retrieval-augmented generation, and fine-tuning. It also evaluates different LLMs, such as OpenAI and Llama 2. Developing evaluation techniques has been a key aspect of this project, alongside assessing Gaby's usefulness, accuracy, and ethics. This research aims to understand how AI can enhance user interactions with library resources while identifying areas for improvement and future development.