DOI: 10.3138/jmbs-2025-0009 ISSN: 2631-8407

The Impact of AI Voice Tutors on Phonetic Acquisition Among Saudi EFL Learners: A Quasi-Experimental Study at Jouf University

Ahmad Mugableh, Mamdouh Alhuwaykim

Abstract

This study investigates the impact of AI-powered voice tutors on the phonetic acquisition of Saudi EFL learners. Although digital technologies are increasingly integrated into language education, empirical research has yet to examine the extent to which AI-driven pronunciation feedback yields measurable phonetic gains in Arab EFL contexts. This quasi-experimental study evaluates whether personalized AI feedback enhances both segmental accuracy (e.g., /p/–/b/, /v/–/f/, /ɪ/–/iː/) and suprasegmental control (e.g., word-stress patterns) among beginner-to-intermediate learners. In this study, 60 Saudi male students (18 to 20 years of age) who were enrolled in Jouf University's English Foundation Program participated in a 6-week intervention. The experimental group engaged with the ELSA Speak application (an AI-based pronunciation tutor that incorporates ASR and acoustic analysis algorithms), while the control group received conventional pronunciation instruction. Pre-test and post-test speech recordings were acoustically analyzed using PRAAT and were supplemented by perception tasks and attitude questionnaires. A paired-sample t- test and an independent-sample t- test revealed statistically significant gains in the AI group ( p  < .001), with large effect sizes across segmental and suprasegmental measures. Acoustic analysis demonstrated increased voice-onset time for /p/, improved formant separation in /iː/–/ɪ/ contrasts, and enhanced stress-timing ratios. Qualitative findings further indicated heightened learner motivation, reduced pronunciation anxiety, and positive perceptions of immediate visual feedback. The results suggest that AI voice tutors can substantially enhance both perception and production of English phonetic features through sustained adaptive feedback, an instructional dimension often constrained in large EFL classrooms. The study provides empirical support for integrating AI-assisted pronunciation systems into Saudi higher education curricula while promoting pedagogically guided collaboration between instructors and intelligent technologies.

More from our Archive