Unlocking Educational Potential: Exploring Students’ Satisfaction and Sustainable Engagement with ChatGPT Using the ECM Model

Aim/Purpose: The main goal of this study is to investigate the factors affecting students’ satisfaction and continuous usage of ChatGPT in an educational context, using the Expectation-Confirmation Model (ECM) as the theoretical framework. Specifically, this investigation focuses on identifying how user expectations, perceived usefulness, and satisfaction influence the continuous usage of ChatGPT in education. Background: ChatGPT is an AI-based chatbot that can generate natural language and support multi-turn dialogue. ChatGPT can be used in education for various purposes, such as academic writing, language translating, coding assistance, and information searching. Research indicates that ChatGPT is highly beneficial for students. Specifically, it aids in time-saving by synthesizing information across diverse fields and offers feedback tailored to individual learning needs and progress. Moreover, it enhances learning experiences through a personalized learning process, as well as aids in the translation of learning materials into different languages for improved accessibility. Despite the potential of ChatGPT in education, there remains a gap in understanding the determinants of user satisfaction and continuous usage. Therefore, this study was conducted to bridge this gap by assessing the impact of expectation confirmation, perceived usefulness, and satisfaction on continuous usage of ChatGPT. Methodology: This research employed a quantitative approach, which utilized a five-point Likert scale questionnaire to measure four variables with a total of 14 items. Data were collected from 435 students from eight Vietnamese universities through non-probability convenience sampling. The internal consistency of the questionnaire was examined using Cronbach’s alpha. Four-factors Confirmatory Factor Analysis (CFA) and Structural Equation Modeling (SEM) were then employed through IBM SPSS 26 and AMOS 24 to assess the proposed model and hypotheses, allowing for a comprehensive examination of the relationships between 4 variables: perceived usefulness, expectation confirmation, satisfaction, and continuous usage. Contribution: This paper contributes to the literature on ChatGPT in education by identifying how expectation confirmation and perceived usefulness affect students’ satisfaction and their intention to continue using ChatGPT in their learning. This study also has practical implications for educators and students for applying ChatGPT in their learning and teachin