Enhancing Oral Communication in Real Time: University Students’ Self-Directed Use of AI-Powered Speech Recognition in English Classrooms

The increasing availability of AI-powered speech recognition has created new possibilities for supporting oral communication in English classrooms. This study explores how university students use AI-powered speech recognition tools in a self-directed manner during real-time speaking activities and how they perceive the usefulness of these tools for enhancing oral communication. Adopting a convergent mixed-methods design, the study integrates qualitative data from classroom observations and semi-structured interviews with quantitative data from a descriptive Likert-scale questionnaire. The qualitative findings reveal that students use speech recognition strategically to monitor intelligibility, rehearse spoken output, and manage speaking-related anxiety during communicative tasks. Rather than replacing interaction, the tool functions as a flexible support that learners draw on selectively according to situational needs. The quantitative results indicate generally positive perceptions of the tool’s usefulness, particularly in relation to confidence, fluency awareness, and ease of use. By combining observed practices with learners’ reported perceptions, this study offers a classroom-grounded account of AI-powered speech recognition use in real-time speaking contexts. The findings contribute to ongoing discussions on AI-assisted language learning by highlighting the pedagogical potential of speech recognition as a learner-oriented resource for supporting oral communication.