A Integrating AI-Powered Chatbots into EFL Speaking Classes: A Mixed-Methods Study on Learners’ Communicative Confidence and Interaction Patterns

Abstract

Artificial intelligence (AI) has emerged as a transformative force in education, reshaping the ways learners engage with language, content, and communication. Within English as a Foreign Language (EFL) pedagogy, the integration of AI-powered chatbots has generated increasing interest due to their ability to provide authentic conversational practice and individualized feedback. This study investigates the impact of AI chatbot integration on learners’ communicative confidence and interactional behavior in EFL speaking classes. Employing a mixed-methods design, the study involved 52 intermediate-level Iranian learners over eight weeks, during which participants interacted regularly with ChatGPT for structured and semi-structured speaking tasks. Quantitative data from pre- and post-intervention questionnaires revealed a statistically significant improvement in self-reported communicative confidence (t(51) = 8.56, p < .001, d = 1.21), while qualitative analyses of interviews and chat transcripts highlighted reductions in speaking anxiety, greater fluency, and improved turn-taking skills. Findings underscore the pedagogical potential of AI-powered tools for promoting learner autonomy, affective engagement, and interactional competence in the EFL context. The study concludes with recommendations for ethical and pedagogically sound integration of chatbots into communicative language teaching frameworks.