Using Gemini Voice-To-Text in Facilitating Pronunciation Awareness Through Self-Correction Among Eighth-Grade Students

This study aims to investigate how Gemini Voice-to-Text facilitates pronunciation awareness through self-correction among eighth-grade students in an English as a Foreign Language (EFL) classroom context. Specifically, it examines students’ experiences using the tool, how AI-generated transcription supports their awareness of pronunciation differences, and the practical benefits and challenges experienced during implementation. This study used a qualitative case study design. The participants in the study were eight eighth-grade students from SMP Negeri 16 Surabaya, East Java, Indonesia. These students participated in pronunciation activities using Gemini Voice-to-Text. Data were collected through classroom observation, semi-structured interviews, and documentation of the AI-generated transcription output. Analysis of the collected data was conducted thematically with data familiarization, coding, categorization, and developing themes. The findings revealed that students used Gemini Voice-to-Text through three interconnected processes: accessing the tool during speaking activities, reviewing AI-generated transcription, and comparing their spoken production with the recognized output. These processes provided students with opportunities to notice pronunciation differences and engage in self-correction through repetition and modification of their pronunciation. Students also perceived the tool as beneficial because it provided immediate feedback, supported independent practice, and increased opportunities for repeated pronunciation practice. On the other hand, there were some problems with using the tool, especially regarding mistakes in the transcription due to varying ways people say words, how fast someone speaks, and the amount of noise in the classroom. This research concludes that Gemini Voice-to-Text can facilitate pronunciation awareness through self-correction when its feedback is critically interpreted by students and supported by teacher guidance.