A Comparative Analysis of Teacher and AI-Supported Corrective Feedback on Fossilized Pronunciation Errors Among EFL Learners

Pronunciation is a core element of communicative competence in the English as a foreign language (EFL) learning process. However, learners at the interlanguage stage often display fossilized errors. Artificial intelligence (AI) offers potential as a corrective feedback tool to detect and address such errors. This study investigates fossilized pronunciation errors among EFL learners, examines the effects of teacher- and AI-supported corrective feedback, and compares their effectiveness. The quasi-experimental design involved 44 EFL learners. Data collection tools included a background questionnaire, a pilot test, a test of mispronounced words, and a focus group interview. Descriptive statistics, the Mann-Whitney U test, and the Wilcoxon Test were applied to analyze the data. Findings revealed that learners in both groups exhibited moderate levels of fossilized pronunciation errors. Teacher feedback was found to reduce such errors and improve overall pronunciation skills, while AI-supported feedback contributed to error reduction for certain words.