Artificial intelligence applications in surgical education and training: a systematic review

Introduction: Surgical training is shifting toward scalable, data-driven education as operative complexity and patient safety expectations increase. AI can support objective feedback and cognitive guidance. Methodology: We conducted a PRISMA-guided systematic review using a PICO framework and searched PubMed, Scopus, and IEEE Xplore for peer-reviewed studies published between 2020 and 2025. Results: Of 1,109 records, 21 studies met the inclusion criteria. Included studies covered simulation-based, robotic, laparoscopic, and computer-assisted training using deep learning/computer vision, tutoring or predictive models, and language-based tools. Performance outcomes predominated (81.0%) over engagement-related outcomes (19.0%), and reported effects were mainly positive (76.2%) or increased (23.8%). Discussion: Evidence suggests near-term, task-specific gains when AI provides objective measurement and feedback, but comparability is limited by heterogeneous endpoints, small samples, and single-center designs. Conclusion: AI-enabled surgical education shows promise for objective assessment and adaptive instruction, but multicenter longitudinal studies with standardized metrics are still needed.