AI Conversational Agent Design for Supporting Learning and Well-Being of University Students

University students often face a myriad of challenges that can adversely affect their learning and overall health. However, re-search has shown that a minimal number of students access sup-port for their learning and well-being problems. In this project, we aim to evaluate the feasibility of using artificial intelligence (AI) chatbots powered by the generative pre-trained transformer (GPT) to support student learning and well-being. We consulted with clinical psychologists and counselors to identify objectives and iteratively develop the AI conversational agent to align closely with real-world scenarios. 63 questions from different domains students might pose are designed to assess the effectiveness and professionalism of the developed AI conversational agent’s re-sponse. The chatbot's performance was evaluated using context-specific metrics. Additionally, students were recruited to engage in conversations with the chatbot, and their feedback was collect-ed for thematic analysis. Our findings suggest that the develop-ment and evaluation of this AI conversational agent present a promising avenue for addressing student well-being concerns within university environments, while there remains a need to advance their capacity for personalized, empathetic interactions to meet users' needs and expectations fully.