Understanding GenAI adoption in EFL higher education: a dual-model study using SDT and TAM

The exponential emergence of generative AI (GenAI) tools, such as ChatGPT, is reshaping language learning in higher education. However, little is known about the psychological and perceptual factors that sustain students’ long-term use of these technologies. This study develops and empirically tests an integrated framework combining Self-Determination Theory (SDT) and the Technology Acceptance Model (TAM) to explain continuance intention and actual usage of GenAI tools in English as a Foreign Language (EFL) learning. Using a descriptive cross-sectional survey design and non-probability convenience sampling, data were collected from 451 Malaysian university students (Generation Z). Partial Least Squares Structural Equation Modelling (PLS-SEM) and PLSpredict were used to examine direct and mediated effects of autonomy, competence, and relatedness on perceived usefulness and ease of use. The findings confirm that both motivational and technological constructs significantly influence continuance intention, which in turn predicts actual behaviour. By including actual usage, the study addresses the intention – action gap in EdTech research. Situated in a Southeast Asian context, the findings offer actionable insights for culturally responsive, learner-centred AI-enhanced EFL programs. This research is relevant to policymakers, instructional designers, and developers seeking to support the adoption of sustainable GenAI in language education.