Amidst the digital transformation of higher education and artificial intelligence, particularly ChatGPT is reshaping student learning. This study evaluates the impact of ChatGPT’s core technological features (conversational ability, content generation, foundational learning, and contextual understanding) on four dimensions of self-directed learning: motivation, attitude, learning organization, and self-assessment. The research utilised a quantitative approach with 400 students at a higher educational institution in Vietnam. The data were collected and analyzed via structural equation modeling (SEM). Findings reveal that all four components positively influence self-learning, with content generation boosting motivation and learning organization, while conversational and foundational features improve learning attitude. Contextual understanding supports reflection and self-assessment. The model shows high reliability and validity. Based on these findings, several recommendations are proposed, including providing guided instructions for the use of ChatGPT, integrating it into feedback and self-assessment, offering training in AI-assisted learning skills, and personalizing learning strategies. This study contributes to clarifying the role of AI in enhancing sustainable self-directed learning capacity among university students.
Thuy et al. (Tue,) studied this question.