Abstract: Although e-learning is considered one of the leading teaching methods in higher education, both learners and instructors face significant challenges owing to reduced social interaction compared with traditional classroom learning. In this study, we explore the leveraging of recent developments in generative artificial intelligence (AI) and create a custom chatbot using retrieval-augmented generation. A research model combining the technology acceptance model and the interactive-constructive-active-passive theory was developed and used to investigate how the chatbot affects students’ perceptions and perceived learning outcomes in online and blended classes. This study provides empirical evidence indicating that custom chatbots can be integrated into higher education to enhance students’ e-learning experiences, and through interacting with chatbots, students’ behaviors shift from passive to interactive engagement. The findings shed light on how generative AI helps to improve e-learning experience, highlighting the effectiveness of such technology in support of social interaction and emotional engagement in higher education. The study also demonstrates the feasibility of deploying custom AI chatbots in college classes and provides practical recommendations.

Keywords: Chatbot, e-Learning, Generative AI in teaching, Higher education, Online learning

Download This Article: JISE2025v36n4pp384-399.pdf

Recommended Citation: Chen, J. M., Zhang, L., Pengnate, S., Ma, E., & Leung, X. Y. (2025). Integrating a Custom Chatbot Into Higher Education: From Passive to Interactive E-Learning. Journal of Information Systems Education, 36(4), 384-399. https://doi.org/10.62273/FISY8338