ChatGPT’s effects on EFL learners’ autonomy and self-efficacy in self-regulated learning contexts

Abstract

This study compares the effects of ChatGPT-facilitated and human teacher-facilitated project-based learning (PBL) on learner autonomy and self-efficacy among Vietnamese EFL students in a language training center. Using an explanatory sequential mixed-methods design, fifty participants were randomly assigned to either a ChatGPT-facilitated group or a teacher-facilitated group for an eight-week project-based intervention. Learner autonomy and self-efficacy were measured before and after the intervention. Quantitative data were analyzed using independent t-tests and one-way ANCOVA, while qualitative data from focus group interviews were coded and thematically analyzed. The results indicated that the ChatGPT group achieved significantly greater gains in learner autonomy, whereas both groups improved equally in self-efficacy. Qualitative analysis revealed that AI facilitation encouraged group initiative and experimentation, while teacher facilitation reinforced established routines and delegated authority. These findings demonstrate the potential of AI-driven facilitation to promote learner autonomy and suggest that blended models combining AI and human strengths may optimize language learning outcomes.