An artificial intelligence-supported GFCA learning model to enhance L2 students’ role-play performance, English speaking and interaction mindset

Abstract Role-play tasks have long been used by researchers and practitioners to observe L2 (Second language) speaking performance. This social-situated simulation allows students to employ their language skills to converse about real-life themes. While role-plays are highly plausible to actively engage students in interactive learning environments, it has been challenging to determine whether students perform at an adequate level of speaking proficiency with an appropriate learning approach. Nevertheless, the emerging technology-supported role-play tasks employing AI (artificial intelligence) could enhance competencies, enabling students to perform better in role-plays. Therefore, to enhance and support students’ performance in role-plays, English speaking skills, and their interaction mindset, we integrated an AI-based speaking learning app in the Generalization, Formulation, Correction, Appreciation (AI-GFCA) learning model. A quasi-experiment was conducted in a university with a total of 45 students. One class was randomly assigned to apply the AI-GFCA learning model as the experimental group, while the other was the control group (AI-C). The findings indicated that the AI-GFCA learning model could significantly enhance students’ role-play performance, English speaking skills, and interaction mindset. Furthermore, students produced fewer L2 errors and perceived a better learning experience than the AI-C group. It is noted that with the support of AI learning speaking through role-play tasks, students received sufficient corrective feedback, which encouraged them to establish a positive and motivating learning interaction, thus benefiting their academic performance.