Vietnamese University Students’ Perceptions of Generative Artificial Intelligence in Higher Education: A Qualitative Study Through the Technology Acceptance Model

Abstract

Background: Generative artificial intelligence (GenAI) has diffused rapidly across higher education worldwide, yet empirical evidence from Southeast Asian and Global South contexts remains limited. Vietnamese higher education, characterised by rapid expansion, strong digital-transformation agendas, and examination-oriented assessment cultures, provides a distinctive setting in which to examine student perceptions of GenAI. Objective: This study examines Vietnamese university students' perceptions and adoption of GenAI in higher education through the Technology Acceptance Model (TAM), with particular attention to phenomena that extend beyond classical TAM constructs. Methods: An exploratory qualitative design was employed. Semi-structured in-depth interviews were conducted in Vietnamese with 30 university students from diverse academic disciplines (business, engineering, social sciences, design, media, and fine arts). Data were analysed thematically following Braun and Clarke (2006), yielding five principal themes and sixteen content codes. Results: GenAI has become a near-ubiquitous learning tool among participants, primarily used for information retrieval, document summarisation, presentation preparation, and completion of academic tasks. High perceived usefulness and perceived ease of use facilitated widespread adoption (90% reported daily or weekly use). Substantial concerns were identified regarding over-reliance, technological dependency, inaccuracies in AI-generated content, diminished independent thinking, and challenges to authentic assessment. A distinct pattern of instructor-side AI overuse - lecturers generating teaching materials without sufficient verification - was also documented. Students in creative disciplines expressed particularly strong scepticism towards the uncritical promotion of AI and concerns about its potential impact on future professional prospects. Conclusion: Although TAM explains the rapid acceptance of GenAI, it does not fully account for concurrent epistemic distrust, AI fatigue, dependency concerns, and educational-quality anxieties. The findings contribute empirical evidence from a Global South context and demonstrate that GenAI adoption in Vietnamese higher education is characterised by a tension between high pragmatic acceptance and pronounced critical reservations. An extended TAM framework incorporating epistemic, pedagogical, and professional dimensions is proposed.