Collaborative AI Prompting as a Driver of Metacognitive Awareness and Epistemic Agency in Higher Education

The growing diffusion of generative AI in higher education has shifted attention toward learners’ capacity to formulate, evaluate, and refine their interactions with AI systems [1, 2]. From a Knowledge Building perspective [3], AI functions as a mediating artefact for collaborative knowledge construction, provided learners exercise epistemic agency, the capacity to identify problems, set inquiry goals, and advance knowledge [4], both individually and collectively [5]. Equally central is metacognitive regulation: planning, monitoring, and evaluating one’s own cognitive processes [6–9] becomes critical when students must assess AI outputs against authoritative sources and revise strategies accordingly. Given limited research on this intersection, the study asks: How do metacognitive regulation and collective epistemic agency emerge during collaborative prompt formulation, evaluation, and revision in AI-supported higher education?