INTEGRATING ARTIFICIAL INTELLIGENCE IN UNDERGRADUATE MEDICAL EDUCATION: A QUALITATIVE EVIDENCE META-SYNTHESIS

Abstract

Background: The accelerated integration of artificial intelligence into healthcare calls for a similarly accelerated change of undergraduate medical education. However, there still exists a significant gap between the pace of technological change and the institutional training of the future doctor in (formal) training centers. Objective: The aim of this qualitative evidence synthesis was to collecte global perspectives from medical students and educators that will help to explain their sense of what they think about, how they teach about, how to approach ethics around, and institutional preparedness for integrating AI within Undergraduate Medical Education. Methods: The Cochrane Qualitative Evidence Synthesis searching strategy was followed throughout a systematic search of PubMed, Scopus, and ERIC from 2018 to 2024. Twenty-five qualitative and mixed methods studies were chosen, using the CASP checklist and critically appraised, and thematically synthesised. Confidence in the results of findings was expressed through the evaluation of CERQual. Results: The analysis provided four key themes: (1) Perceptions and Preparedness, including student curiosity as well as apprehension at the potential role displacement; (2) Pedagogical Shifts, illustrating a need for integrated, adaptive learning tools in place of individual courses; (3) Ethical Dimensions, reflecting issues of bias, accountability and data privacy; and (4) Institutional Readiness, including faculty unpreparedness as well as a lack of resources identified as key factors. The quality of evidence as high certainty in the areas of student demand and ethical issues and moderate certainty in implementation strategies were identified using CERQual. Conclusion: The successful use of artificial intelligence during the medical education process should involve a proportional mixture of pedagogical innovations, training of the faculty as a whole, and ethical standards. In summary, our task in the future is to make an effort to design cooperative and context aware AI models that facilitate the clinical judgment without jeopardizing the human element of medicine.