Effects of incorporating AI chatbots into think–pair–share activities on EFL speaking anxiety, language enjoyment, and speaking performance

A major challenge for English as a foreign language (EFL) students is having limited opportunities to practice speaking in conducive environments, and this limitation can affect their speaking performance. Emotions such as speaking anxiety and language enjoyment can influence EFL students’ achievement. The present study investigated the effects of incorporating an artificial intelligence (AI) chatbot into think–pair–share (TPS) activities on foreign language speaking anxiety (FLSA), foreign language enjoyment (FLE), and speaking performance. We used a pretest–posttest control and a quasi-experimental design and conducted experiments over 6 weeks. The participants comprised 75 first-year students selected from a state university in Indonesia during the fall semester of the 2022–2023 academic year; 39 students were assigned to the experimental group, engaging in TPS activities with an AI chatbot, and 36 were assigned to the control group, engaging in conventional TPS activities. Data were collected using an 18-item FLSA questionnaire, a 10-item FLE questionnaire, and a speaking test based on the IELTS speaking evaluation rubric. The results revealed that the experimental group significantly outperformed the control group. Incorporating the AI chatbot into TPS activities notably reduced the students’ FLSA, enhanced their FLE, and improved their speaking performance. Moreover, interviews with five randomly selected students revealed positive perceptions regarding the use of AI chatbot in TPS activities, supporting the quantitative results. Future research should consider longer observation periods and comprehensive evaluations to further investigate the effect of AI chatbots in TPS activities across cognitive and affective domains.