Differences in perceived influences and educational support needs of generative AI among medical and nursing students across educational levels: a descriptive qualitative study

Abstract

Generative artificial intelligence (GAI) has been extensively integrated into nursing and medical education, significantly transforming the academic practices of students. Nonetheless, the varying impacts of GAI on core competency of student across educational stages remain ambiguous, with limited empirical evidence on hierarchical educational support needs. Drawing on cross-sectional comparisons of GAI-related experiences among undergraduate, master’s, and doctoral students, this study aimed to explore differences in students’ perceptions of GAI influences across educational levels. It focuses on students’ perceived impacts concerning critical thinking, research competence, and academic integrity, and further identify personalized support needs for GAI-oriented education at each stage. This cross-educational-stage comparative perspective is expected to provide empirical evidence for constructing an AI-literacy education system tailored to learners at different levels. This descriptive qualitative study draw on semi‑structured interviews with 44 students (16 undergraduate, 16 master’s, 12 doctoral) from a western Chinese medical university using inductive content analysis, with Technology Acceptance Model (TAM) and AI Literacy frameworks informing discussion. Interview findings showed that GAI-use patterns varied across academic levels: undergraduates for coursework, versus postgraduates for literature reviews, experimental design, and data analysis. Students perceived that GAI-use could yield benefits for their critical thinking and knowledge acquisition, counterbalanced by risks of over-reliance, academic integrity, and cognitive offloading—with research competency gains largely confined to postgraduates. A pronounced demand for systemic support, including teacher guidance and institutional training, pervaded all levels. Interview findings suggest that GAI may exert differentiated perceived influences on the development of medical and nursing students. Students’ GAI usage patterns, academic benefits and potential risks may vary across academic levels, which suggests the necessity of adaptive educational support tailored to distinct stages. Based on qualitative findings, we propose an extended three level TAM model (PEOU driven, PU driven, critical reflection) as a preliminary conceptual framework to capture students’ different GAI use patterns. The findings may inform stratified AI literacy education and stage specific curriculum in health professions education, though further validation of this model in diverse educational contexts is still required.