Students' Continuance Intention to Use Generative Artificial Intelligence forAcademic Purposes: An Extended Expectation-Confirmation Model from aVietnamese University

Abstract

This study investigated why students continue using generative artificial intelligence (Generative AI) for academic purposes at the Interlingual Institute of Education, Nguyen Tat Thanh University. It extended the Expectation Confirmation Model with two SDT-informed perceptions (autonomy and competence) and two Generative AI characteristics (perceived intelligence and anthropomorphism). Survey data from 269 students with prior academic experience using Generative AI were analyzed using PLS-SEM. Results have shown that seven direct relationships were supported, while the usefulness – satisfaction path was marginal and the competence – usefulness path was not significant. The model explained 37.3% of the variance in continuance intention. Perceived usefulness was the strongest direct predictor, followed by satisfaction. The findings extended post-adoption research in an underexplored Vietnamese international higher education setting and suggested that universities should strengthen the academic value, user satisfaction, learner autonomy, and responsible use of Generative AI.