THE EFFECT OF USING ELEVENLABS PLATFORM BASED ON AI AUDIO TOWARDS STUDENTS’ LISTENING SKILL

Abstract

This study investigated whether ElevenLabs AI-generated audio significantly affected students’ listening skills and described its implementation in listening instruction. A quantitative pre-experimental one-group pre-test–post-test design was employed. The population comprised 35 tenth-grade students at SMK NU AL-YAMAN in the 2025/2026 academic year. Using purposive sampling, 20 students of Class X TKJ 1 were selected as the experimental group, while 15 students of Class X TKR 1 participated in the instrument try-out. Data were collected through a 25-item listening test and an observation checklist completed by the English teacher. The participants completed a pre-test, received six treatment sessions using recount-text audio generated through ElevenLabs, and completed a post-test. The test scores were analyzed with SPSS using the Shapiro–Wilk normality test and a paired-samples t-test, while the observation data were summarized descriptively and supported by thematic coding. The mean score increased from 12.75 on the pre-test to 21.65 on the post-test. The paired-samples t-test yielded t(19) = −14.175, p < .001, indicating a statistically significant improvement after treatment. Ten of twelve observation indicators were achieved, showing that the audio was implemented through pre-listening, while-listening, and post-listening activities and that students generally demonstrated attention, participation, and interest. ElevenLabs AI-generated audio therefore supported a significant within-group improvement in listening performance. Nevertheless, because the design did not include a control group, the result should not be interpreted as conclusive evidence that ElevenLabs alone caused the improvement.