Pemanfaatan Teknologi Speech-to-Text untuk Penilaian Diri dalam Pengucapan Bahasa Inggris bagi Pembelajar EFL

Abstract

The latest development in Automatic Speech Recognition (ASR) technology has revolutionized speaking practice for English as a Foreign Language (EFL) students. Furthermore, based on the systematical review of literatures containing 25 published peer-reviewed research articles ranging from 2020 to 2025, has been examined the effectiveness of speech-to-text tools in facilitating self-assessment of speaking practice. The studies choosen and analysed by using mixed methods which focused on qualitative with PRISMA Guide. The results reveal 3 central points, they are: 1). ASR tools are significantly improving students’ speaking accuracy, especially on vowel sounds and regular past tense suffix; 2). The tools give visual and objective direct feedbacks that improve autonomous and motivation of learning; 3). Althouh effective, ASR system still facing challenge on recognizing non-standard accents and requires optimum audio condition for better performance. Recent evidence also shows that regular and structured ASR practice supports longterm retention to the improvement of speaking skill, though the suprasegmental features like intonation and stress still feel difficult to master. On the other hand, this article also offers practical recommendation to integrate the technology in language curriculum, as well as suggestion for future research on improving accent introduction and the implementation in real world context. By mediating technology innovation and evidence-based pedagogy, this study provides practical knowledge for teachers, curriculum designers, and researchers who eager to implement ASR as a motivating and sutainable speaking practice tool in EFL context which based on either digital or even blended learning.