EC Neurology

Introduction: Artificial intelligence (AI) is being integrated into medical education at an accelerating pace, yet the empirical evidence supporting its effectiveness remains heterogeneous and insufficiently synthesized. Much of the strongest quantitative evidence currently available concerns simulation-based education broadly rather than AI-specific applications, and this distinction is maintained throughout the review.

Objective: To critically synthesize and compare quantitative evidence on different AI applications in medical education, with emphasis on effect sizes, study quality, and the specificity of evidence to AI (as opposed to simulation technology in general).

Methods: Narrative review of peer-reviewed literature from PubMed and Scopus (2019-2026), prioritizing systematic reviews and meta-analyses that reported effect sizes or objective outcomes, supplemented by recent primary studies where AI-specific evidence was otherwise unavailable. This is a narrative, not a systematic, review: searches were not exhaustive, dual-reviewer screening was not performed, and no formal quality-appraisal instrument (e.g. MERSQI) was applied.

Results: Foundational meta-analyses of technology-enhanced simulation (not AI-specific) show large effects on knowledge and skills versus no intervention (Hedges' g ≈ 1.1-1.2) and small-to-moderate effects versus traditional instruction (g ≈ 0.30-0.66). Virtual patient meta-analyses show low-to-moderate positive effects on clinical reasoning and skills; a 2025 crossover study specifically comparing an AI-enhanced virtual patient platform to a conventional one found higher learner ratings across several design domains. Evidence for AI-driven adaptive learning and automated assessment tools is more limited, derived mainly from single-institution studies with non-standardized outcome measures (e.g. Cohen's d = 0.78 for study engagement in one adaptive-learning trial; up to 87.5% grading-time reduction in one automated-assessment study). Significant gaps remain in long-term clinical outcomes and equity research.

Conclusion: Simulation environments have the most robust supporting evidence, though most of that evidence predates and is not specific to AI. Evidence specifically attributable to AI is strongest for virtual patient applications and weakest for adaptive learning systems, which remain at an early, single-site stage of evaluation. Higher-quality, longitudinal, AI-specific research is needed.

Keywords: Artificial Intelligence; Medical Education; Virtual Patients; Simulation; Adaptive Learning; Narrative Review

1. Nasca TJ., et al. “The next GME accreditation system - rationale and benefits”. New England Journal of Medicine 11 (2012): 1051-1056.
2. Gordon M., et al. “A scoping review of artificial intelligence in medical education: BEME Guide No. 84”. Medical Teacher 4 (2024): 446-470.
3. Cook DA., et al. “Technology-enhanced simulation for health professions education: a systematic review and meta-analysis”. Journal of the American Medical Association 9 (2011): 978-988.
4. Cook DA., et al. “Comparative effectiveness of technology-enhanced simulation versus other instructional methods: a systematic review and meta-analysis”. Simulation in Healthcare 5 (2012): 308-320.
5. Kononowicz AA., et al. “Virtual patient simulations in health professions education: systematic review and meta-analysis by the Digital Health Education Collaboration”. Journal of Medical Internet Research 7 (2019): e14676.
6. Borg A., et al. “AI-enhanced social robotic versus computer-based virtual patients for clinical reasoning training in medical education: observational crossover cohort study”. Journal of Medical Internet Research 27 (2025): e82541.
7. Y Chen. “Evaluation of the impact of an AI-driven personalized learning platform on medical students' learning performance”. Frontiers in Medicine (Lausanne) 12 (2025): 1610012.
8. Campbell KK., et al. “Applying state-of-the-art artificial intelligence to grading in simulation-based education: assessment, feedback, and ROI”. Discover Artificial Intelligence 5 (2025): 202.
9. McGaghie WC., et al. “Does simulation-based medical education with deliberate practice yield better results than traditional clinical education? A meta-analytic comparative review of the evidence”. Academic Medicine 6 (2011): 706-711.
10. Association of American Medical Colleges. “Principles for the responsible use of artificial intelligence in and for medical education”. Washington, DC: AAMC (2025).