Learning to question the machine: Prompt literacy, self-regulation, and AI-assisted interpretation in an EFL mythology classroom

This study examines how structured prompt literacy instruction shaped students’ use of ChatGPT in a humanities-based English as a Foreign Language (EFL) mythology course. Drawing on self-regulated learning (SRL) as the analytic frame, and using situated learning and GenAI literacy as lenses, the study explores how students used AI as coach and mentor during a 15-week intervention. Data consisted of reflective journals, semi-structured interviews, and classroom observations from 21 undergraduates and were analyzed through reflexive thematic analysis. Five themes were generated: interpretive shift through metacognitive monitoring, prompt development as learner agency, conditional epistemic trust, differentiated uses of coaching and mentoring, and the continued centrality of the instructor. The findings suggest that prompt literacy functioned as a metacognitively oriented practice through which students planned, monitored, and revised AI-assisted inquiry. Students reported more deliberate questioning, stronger evaluation of AI output, and greater willingness to verify responses against course materials. They preferred coaching before mentoring, particularly when tasks were unfamiliar or demanding. The study does not claim causal effects, but it shows how explicit prompt instruction can support reflective AI use in EFL teaching. Pedagogically, the study supports prompt literacy instruction, coaching before mentoring, and instructor guidance as essential to disciplinary judgment.