Exploring University Students’ Engagement with ASR-Mediated Speaking Practice through Individual Digital Portfolios

This qualitative action research explored university students’ engagement with ASR(automatic speech recognition)-mediated English speaking practice through individual digital portfolios in a general English course. Building on a previous instructional cycle that relied on a shared Padlet, the study emphasized individual digital portfolios while maintaining the use of a limited shared Padlet for reference. Data were drawn from semi-structured interviews with five focal participants and digital portfolio artifacts and were analyzed using inductive thematic analysis. Findings indicate that learners’ engagement with ASR-mediated speaking practice developed gradually through repeated practice, accompanied by a reduction in anxiety and the routinization of speaking tasks. ASR transcription feedback supported learners’ noticing of pronunciation-related mismatches and subsequent revision, but interview data revealed a tendency to over-prioritize ASR recognition accuracy over communicative naturalness. Instructor feedback played an important role in sustaining learner confidence and continued participation. The study underscores the importance of pedagogical mediation and flexible portfolio design in ASR-mediated speaking instruction, particularly in supporting learners’ reflective engagement over time.