Learning to Speak English with AI-Based Speech Recognition: An Autonomous Learning for EMI Students

Abstract

Learning to speak English fluently is widely recognized as one of the most challenging aspects of language learning. AI tools provide a low-pressure environment for practice, which enhances fluency and learner confidence, crucial for EMI students’ speaking performance. Therefore, this study intends to describe how students feel about using AI-based speech recognition to enhance speaking skills. Besides, it is to outline the challenge faced by students when using AI-based speech recognition as autonomous learning tools. This study employed qualitative methodology and descriptive research design. Ten students of Arabic Education of Islamic University in Central Java, Indonesia in the academic year 2025/2026 participated in this study. They were students of International Class Program that used English as a Medium of Instruction in teaching learning process. Interviews and documentation were used as data gathering methods. Data reduction, data display, and conclusion drawing or verification were done using data analysis techniques. Based on the findings, students view AI-based speech recognition as tools for autonomous learning to support their learning in speaking outside the classroom. Further, it can help them improve their vocabulary, pronunciation, and grammar as well as confidence in speaking and demonstrate enhanced quality of discussion during classroom activity. However, the use of AI-based speech recognition has some challenges, such as limited free features, insufficient offline capability, and an excessive dependency on instant feedback and internet limited access.