Determinants of Learners’ Experience in Using AI Chatbots

Beh, Mun Hui (2025) Determinants of Learners’ Experience in Using AI Chatbots. Masters thesis, Tunku Abdul Rahman University of Technology and Management.

Text 04 Beh Mun Hui.pdf Restricted to Registered users only Download (3MB)

Abstract

Purpose: This study explores the factors that influence student learning engagement when using AI chatbots in higher education. It examines the direct and mediated effects of University Administration-Student Relationship (UASR) and Dean’s Office-Student Relationship (DOSR), and User Experience (UEX) on Learning Engagement (LE), with Technostress (T) acting as a moderator. Design/Methodology/Approach: A quantitative research design was used. This study utilised a structured online questionnaire distributed to university students in Malaysia. Data from 387 respondents were analyzed using SPSS version 25.0 and Partial Least Squares Structural Equation Modeling (PLS-SEM) through SmartPLS 4. The reliability, validity, and predictive strength of the model were all confirmed. Findings: The results indicate that UASR, DOSR and UEX have significant positive direct effects on LE. UEX serves as a significant mediator between both UASR-LE and DOSR-LE. Furthermore, contrary to the initial hypothesis, technostress (T) was found to have a positive moderating effect while simultaneously maintaining a direct negative effect on learning engagement (LE). This indicates that a high-quality user experience can contextualize technostress, potentially making it a motivating factor. In the absence of good user experience, technostress directly reduces student engagement. Practical Implications: Universities should prioritize strengthening institutional communication and support systems to enhance trust in AI tools. Furthermore, investing in the design of intuitive, user-friendly chatbot interfaces is crucial to mitigating the negative effects of technostress and leveraging its potential motivating influence. Originality/Value: This research extends the Chatbot-Human Interaction Satisfaction Model (CHISM) by integrating institutional support dimensions and technostress, offering a more holistic framework for understanding AI chatbot adoption in education. The findings provide actionable insights for educational institutions to design and implement AI chatbots that are both effective and psychologically attuned to student needs.