STUDENTS’ PERCEPTION ON THE USE OF SHADOWING TECHNIQUE USING TEXT-TO-SPEECH AI TO IMPROVE ENGLISH PRONUNCIATION

Pronouncing English words correctly has become a familiar struggle experienced by Indonesian EFL students in the classroom. In addressing the issue, researchers employed an adapted procedure of shadowing technique. Here we substituted ready-made recording speech with Text-to-Speech AI generated voice as an inventive method. A case study approach was employed by involving semi-structured interviews, interviewing two welding engineer students as high achiever and mid achiever as the participant, both enrolled in Vocational High School 1 Singosari. The result indicates a pronunciation improvement to some extent, mainly awareness of phonemes, rising and falling pitch, and students were more fluent than before in a way of speaking confidence. Students find the idea straightforward, flexible, and adaptable where they were able to simply generate a voice made from desired script, furthermore they could mimic the generated voice easily, though students might find it difficult to catch the AI speaker in doing complete shadowing. Despite many positive results, the study did not go as planned due to time constraints and participants availability. This finding indicates the necessity for future research to be more meticulously designed and executed.