Artificial Intelligence in English Phonology: Improving Pronunciation and Speech Recognition for Iraqi Universities' Students

Abstract

This study investigates the role of Artificial Intelligence (AI) applications in improving English phonology, pronunciation, and speech recognition among Iraqi universities' learners of English as a Foreign Language (EFL), since despite significant emphasis on grammar and vocabulary in Iraqi universities' higher education, phonological competence remains underdeveloped and continues to hinder oral communication and intelligibility· The research specifically focuses on undergraduate students from the University of Baghdad, Al-Mustansiriyah University, and Al-Mansour University College, where persistent difficulties with segmental sounds such as /θ/, /ð/, and /ʒ/, as well as suprasegmental features including stress, rhythm, and intonation, have been documented· Using a quasi-experimental design, the study integrates AI-powered tools such as Google Speech Recognition, ELSA Speak, and YouGlish to enhance learners’ exposure to authentic speech models, provide real-time feedback, and develop pronunciation accuracy, with data collected from 120 students (60 in the experimental group and 60 in the control group) over a 12-week intervention period· Pre-test and post-test assessments, conducted through acoustic analysis with Praat software alongside expert phonological evaluation, measured improvements in vowel and consonant production, prosodic accuracy, and overall speech intelligibility· Findings indicate that the experimental group demonstrated significant progress in reducing common phonological errors, particularly in the substitution of /p/ with /b/, the devoicing of final consonants, and the misplacement of primary stress in polysyllabic words, while the integration of AI tools also enhanced learner motivation and self-confidence, lowering the anxiety often associated with oral communication in English·