Nisa Lailatul Qodriyah
SMP Ibnu Abbas Cileunyi Bandung

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

Evaluating the Accuracy of AI-Based Pronunciation Assessment: a Comparative Study with Human Raters in an EFL Indonesian Context Siti Masitoh; Nisa Lailatul Qodriyah
Journal of English Education and Teacher Trainer Vol. 3 No. 2 (2026): Journal of English Education and Teacher Trainer
Publisher : UIN Sunan Gunung Djati Bandung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.15575/educater.v3i2.3911

Abstract

The use of Artificial Intelligence (AI) in English language learning has expanded rapidly, particularly in the area of pronunciation assessment. There are enormous AI-powered applications that support the automated feedback and scoring features. It ensures efficiency and objectivity, especially for teachers. This study conducting a comparative study of the AI assessment results with expert assessors in the context of English as a foreign language. This study utilizes a quantitative approach involving 13 Indonesian EFL students who completed pronunciation tasks assessed both by the AI-ELSA Speak application and by an expert who adapted a rubric. The data were analyzed using the Wilcoxon Signed-Rank test to identify score consistency and discrepancies. Moreover, a descriptive document analysis of ELSA’s scoring display was conducted to explore the extent to which the AI feedback addresses pronunciation features. The results indicate a general alignment between AI and human scores, supporting the potential of AI as a useful tool in pronunciation accuracy. In conclusion, AI highly supports the pronunciation accuracy assessment, but it cannot replace human raters in identifying other aspects of pronunciation assessment, especially in stress, intonation, and rhythm. Nevertheless, it serves as a valuable complementary tool to assess students’ pronunciation as practically as possible.