The Role of AI Practice in Enhancing Vocational High School EFL Learners' Speaking Self-Efficacy in Yogyakarta

Abstract

Research on the use of artificial intelligence (AI) in language learning has grown rapidly in recent years. Most existing studies focus on higher education and involve multiple intervention sessions. However, empirical evidence regarding the impact of a single-session AI-based intervention on students’ speaking self-efficacy remains very limited. This study examines how AI-assisted speaking practice affects the speaking self-efficacy of EFL students at a vocational high school in Yogyakarta. Using a one-group pretest-posttest pre-experimental design, the study involved 63 students selected through convenience sampling. Data were collected via a speaking self-efficacy questionnaire administered before and after structured speaking practice sessions using ChatGPT, then analyzed using a paired-sample t-test. The results indicate that students demonstrated high levels of speaking self-efficacy across all four dimensions: linguistic self-efficacy, self-regulatory, delivery, and performance. The most significant improvement occurred in the performance self-efficacy dimension, while the smallest improvement was found in the linguistic dimension. These findings indicate that AI has the potential to boost students’ selfefficacy in speaking English and can serve as a valuable tool to support the psychological well-being of EFL learners.