Answers are grounded in this paper and cite their sources.
Background: Artificial intelligence (AI) is increasingly integrated into medical-imaging education to support image interpretation, personalised learning, feedback, simulation, and competency development. However, evidence regarding its educational effectiveness and implementation remains heterogeneous. Objective: This systematic review evaluated AI-assisted educational interventions in medical-imaging education, focusing on learning outcomes, diagnostic competencies, and learner-reported and implementation outcomes. Methods: Electronic databases and supplementary sources were searched for studies evaluating learner-facing AI-assisted educational interventions across medical-imaging disciplines. Eligible controlled, uncontrolled, single-group, and within-participant intervention studies were included. Findings were synthesised narratively because of substantial heterogeneity in learner populations, AI technologies, study designs, interventions, comparators, and outcomes. Objective effectiveness outcomes were considered separately from learner-reported and implementation outcomes. Results: Seventeen unique studies met the eligibility criteria. Controlled studies generally reported favourable effects of AI-assisted approaches on image interpretation, diagnostic performance, personalised learning, and feedback. Uncontrolled studies also reported improvements in knowledge or diagnostic competency, although causal interpretation was limited by methodological limitations. Learner-reported and implementation findings were generally favourable, particularly for confidence, perceived knowledge, usefulness, and acceptability. Important concerns included hallucinations, reliability, technological overreliance, accountability, data security, and clinical fidelity. Meta-analysis was not undertaken because of substantial heterogeneity. Conclusion: AI-assisted educational interventions show promise for improving selected learning and diagnostic outcomes in medical-imaging education. However, heterogeneous methods, small samples, uncontrolled designs, short follow-up, and reliance on self-reported outcomes limit certainty. AI should therefore complement, rather than replace, structured educator-led instruction.
No takes yet. Share an insight, caveat, or question.
Lauron et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: