ChatGPT Voice and Role-Play Activities on English Oral Communicative Competence: A Pilot Study with Secondary Learners

Resumen

The rapid implementation of AI (Artificial Intelligence) technology has gained the attention of many educators to enhance complex EFL skills, such as speaking, especially in contexts with limited opportunities for language practice. This pilot study inspects the effects of the use of ChatGPT's voice feature to enhance oral presentations (role-plays). A group of secondary EFL learners in a public school participated in a four-week intervention, engaging in daily practice with ChatGPT voice. In each English session, participants engaged in a role-play practice with the AI feature. A speaking rubric with five sub-skills was designed and used to assess students' performance each week, and a semi-structured interview was conducted to collect specific participants’ perceptions. ChatGPT was employed to evaluate student performance, with the exception of interaction/expressiveness, to get more objective results. To examine students' performance across the four weeks and to address the first research question regarding their progression in oral communicative competence, a repeated-measures analysis of variance (ANOVA) was conducted to observe statistically significant changes over time. Qualitative data were analyzed using thematic analysis. Quantitative data indicated that ChatGPT Voice effectively enhanced EFL students´ oral skills in combination with role-plays; however, no significant differences in gains were observed among subskills. On the other hand, qualitative analysis portrayed positive perceptions of the AI tool for speaking practice in alignment with previous research and major considerations of its usage. The study offers important insights for language teachers and researchers into the use of AI technologies and how they complement traditional instruction and assessment.