<![CDATA[This study investigates the effectiveness of an AI-powered pronunciation tool, ELSA Speak, in enhancing speaking accuracy among EFL learners in Indonesian secondary schools, where pronunciation instruction remains underexplored despite its critical role in communicative competence. The research aimed to examine whether the use of ELSA Speak significantly improves students’ pronunciation accuracy compared to conventional instruction. The participants were 60 ninth-grade students of MTs Swasta Muhammadiyah 22 Padangsidimpuan, divided equally into an experimental group (n = 30) and a control group (n = 30) through purposive sampling. Data were collected through pre-test and post-test speaking assessments evaluating pronunciation accuracy, word stress, intonation, and clarity of speech using a validated pronunciation rubric with inter-rater scoring. The results of an independent samples t-test revealed a statistically significant difference between the two groups (t(58) = 5.214, p = 0.000), with the experimental group achieving a mean post-test score of 80.30 (SD = 6.14) compared to 70.50 (SD = 5.89) for the control group. The large effect size (Cohen’s d = 1.57) further confirms the practical significance of the findings. These findings indicate that the integration of AI-powered pronunciation tools can effectively support English language teaching by providing immediate feedback and opportunities for autonomous practice, with important implications for technology-enhanced language learning in secondary school contexts.]]>
Copyrights © 2026
