Purpose Artificial intelligence (AI)-based learning tools are increasingly integrated into higher education, offering efficiency and instructional support while raising concerns regarding students’ cognitive engagement and ethical risks. This study examines the relationships between students’ dependence on AI-based learning tools and three key outcomes: decision-making autonomy loss, learning passivity, and perceived privacy and security concerns. Design/methodology/approach Using a quantitative research design, data were collected from 300 university students in Vietnam and analysed using structural equation modelling (SEM). The results indicate that greater dependence on AI-based learning tools is positively associated with all three outcomes. Specifically, students who rely more heavily on AI report reduced confidence in independent decision-making, lower levels of active learning engagement, and heightened concern regarding data privacy and system security. In addition, multi-group analysis reveals disciplinary differences, with the relationship between AI dependence and decision-making autonomy loss being significantly stronger among information technology students than among students in business and foreign language disciplines. Findings The findings highlight that the risks associated with AI in education are not inherent to the technology itself but are shaped by patterns of dependence and contextual use. The study contributes to emerging debates on responsible AI adoption in higher education and offers implications for educators and policymakers seeking to balance AI support with student autonomy, engagement, and ethical governance. Originality/value This is one of the first studies examining how artificial intelligence (AI) affects Vietnamese university students' decision-making autonomy loss, learning passivity, and perceived privacy and security concerns.
Le et al. (Tue,) studied this question.