The growing integration of artificial intelligence (AI) into English language education has created new opportunities to enhance reading instruction. This study investigated the effect of AI-driven reading instruction on EFL learners’ reading autonomy and explored learners’ perceptions of AI-driven reading environments. A sequential explanatory mixed-methods design including a quasi-experimental pretest–posttest nonequivalent control-group design followed by the qualitative phase using extreme (polar) case sampling and semi-structured interviews, was employed during a five-week intervention. Fifty-four intermediate EFL learners from a Language school in Kerman, Iran were selected through Oxford Placement Test and were randomly assigned to an experimental group (n = 27), which received AI-driven reading instruction through ReadTheory, and a control group (n = 27), which received traditional reading instruction. Quantitative data were collected using pre- and post-tests of reading autonomy questionnaire, while qualitative data were obtained through semi-structured interviews. The quantitative data were analyzed using Analysis of Covariance (ANCOVA) after checking the required assumptions, and the interview data were analyzed through inductive thematic analysis. The findings revealed that AI-driven instruction significantly enhanced learners’ overall reading autonomy as well as its cognitive, metacognitive, affective-motivational, and action-oriented dimensions. However, no statistically significant improvement was found in the social dimension of reading autonomy. Furthermore, qualitative findings indicated that learners perceived ReadTheory as a tutor-like support system, providing personalized and adaptive learning for them. The study concludes that AI-driven platforms can serve as a valuable pedagogical tool for enhancing learners’ reading autonomy when integrated with appropriate instructional support.
