Abstract
Abstract
The aim of this narrative review was to explore whether artificial intelligence (AI)-based decision tools and clinical decision support systems (CDSS) improve diagnostic accuracy and clinical reasoning in senior medical students compared to traditional diagnostic methods. A synthesis of eight empirical studies published from 2020 to 2025 was found through PubMed and Google Scholar. AI-supported tools brought consistent benefits for diagnostic accuracy in standardised, imaging-dependent tasks. Systems with guided reflection and structured feedback supported diagnostic confidence and reflective reasoning. However, evidence on broader clinical reasoning was mixed. Effectiveness was mostly dependent on the quality of the inputs generated by the students and the students’ ability to critically appraise the system recommendations for the less structured, information intensive clinical tasks. Some studies reported no diagnostic benefit and possible adverse effects including automation bias, decreased diagnostic efficiency and decreased trust in either the system or students own judgement. Overall, AI and CDSS tools seem to work best as adjuncts to, not replacements for, clinical reasoning. The educational value of them depends not only on the technological ability but also on a conscious pedagogical integration, which promotes reflection, critical evaluation and a proper use of the recommendations provided by the system. Additional longitudinal and mixed methods research is needed to assess their long-term impact on clinical reasoning and diagnostic performance.