Understanding how medical students learn in the era of artificial intelligence: a mixed methods study

Abstract

BACKGROUND: As medical education evolves, current teaching practices often remain misaligned with how today's digitally native students prefer to learn. While the use of digital tools is widespread, there is limited clarity on students' learning behaviors, particularly their preferences for self-paced, assessment-driven, and technology-supported strategies. This study explores these patterns using a mixed-methods approach to inform more responsive medical curricula. METHODS: A mixed-methods, cross-sectional study was conducted among undergraduate medical students (n = 432) from three universities in the UAE and Jordan. A 23-item questionnaire, developed through literature review and expert validation, included both quantitative and qualitative components. Exploratory and confirmatory factor analyses (EFA and CFA) were used to establish construct validity. Free-text responses were analyzed using thematic analysis to complement and contextualize the quantitative findings. RESULTS: CFA supported a five-factor, 17-item structure with good model fit (χ² = 180.02, df = 102, χ²/df = 1.77, CFI = 0.97, RMSEA = 0.04, SRMR = 0.053). The identified dimensions were: self-paced learning, exam-oriented learning, partnership in learning, collaborative learning, and AI-enhanced learning. Thematic analysis of 218 qualitative responses revealed eight key themes: flexible learning options, enhanced formative assessment, active teaching, study skills development, collaborative learning, use of technology, clinically focused learning, and resource accessibility. These qualitative themes reinforced and expanded upon the quantitative constructs. CONCLUSIONS: Medical students in the AI era adopt a complex, multidimensional approach to learning that is personalized, flexible, and technology driven. The convergence of quantitative and qualitative data underscores the urgent need to align curricula with students' preferences by promoting self-regulated, interactive, and AI-enhanced learning environments. These findings have critical implications for faculty development, curriculum reform, teaching, student assessment, and the future of learner-centered medical education.

Ähnliche Arbeiten