ChatGPT as tutor and mentor in English phonetics: EFL students’ perceptions of concept learning and engagement

This qualitative study examined EFL students’ perceptions of ChatGPT during a short, structured implementation in an undergraduate English Phonetics course. The focus was conceptual understanding and engagement rather than pronunciation performance. Using exported conversation logs from six required sessions, weekly reflective journals, and post-intervention semi-structured interviews with 35 students, we conducted thematic analysis. Students described two complementary functions. In an AI-Tutor role, ChatGPT was perceived to support concept learning through guided questioning, counterexamples, and criterion-based application, especially for phoneme definition work and minimal pair reasoning. In an AI-Mentor role, students reported support for self-regulated learning, including identifying recurring difficulties, planning study actions, persisting with challenging topics, and verifying outputs against course materials. Students also emphasized that structured prompts helped maintain topical focus and reduced drift. Trust was conditional and was managed through routine checking and instructor oversight. The findings suggest that perceived usefulness depends on instructional design that frames roles, elicits learner reasoning, and embeds verification routines.