A phenomenographic analysis of undergraduates’ conceptions of generative AI-mediated learning

This phenomenographic study identified six qualitatively distinct conceptions of GenAI-mediated learning among 15 Taiwanese undergraduates through interviews: (1) answer acquisition, (2) cognitive offloading, (3) future competency development, (4) question-driven exploration, (5) critical judgment engagement, and (6) collaborative creativity. Notably, “answer acquisition” emerged as the most common conception. These conceptions form a hierarchical spectrum, varying along dimensions of interaction mode with GenAI, perceived GenAI roles, and learner agency, and illustrate a progression from instrumental use toward collaborative knowledge construction. Understanding these variations could inform pedagogical strategies that encourage learners to move beyond surface-level engagement. Consequently, educators could promote deeper learning through critical reflection and collaborative co-construction, which are essential for supporting the responsible and transformative use of GenAI in educational contexts.