Research Summary

Can AI-Enhanced Expert Feedback Outperform Automated Tutoring in Surgical Simulation Training?

Kate Young

Key Highlights:

  • AI-augmented personalized expert instruction significantly improved surgical performance compared with intelligent tutoring alone.
  • Participants receiving personalized AI-informed feedback demonstrated higher performance scores, including in realistic complex tasks.
  • Human educator input enhanced learning outcomes beyond automated tutoring, reducing errors such as bleeding and injury risk.

In a randomized clinical trial of 88 medical students, trainees who received artificial intelligence–informed personalized expert feedback demonstrated superior surgical performance compared with those tutored solely by an intelligent tutoring system. The intervention improved skill transfer to more complex, realistic scenarios and led to better performance metrics, including reductions in bleeding and injury risk.

Surgical education increasingly incorporates virtual reality simulation and artificial intelligence tools, yet the optimal balance between human instruction and automated systems remains unclear. While intelligent tutors offer standardized guidance, the capacity for nuanced, individualized feedback is limited. This study was designed to test whether combining AI-generated performance data with personalized human feedback enhances learning outcomes compared with intelligent tutoring alone.

The trial enrolled medical students at the McGill Neurosurgical Simulation and Artificial Intelligence Learning Centre, all of whom were novice users of the NeuroVR surgical simulator. Participants were randomly assigned to one of three groups: intelligent tutoring only (control), standardized expert feedback mirroring the intelligent tutor, or expert instruction tailored using AI-derived error data. Performance was evaluated through AI-calculated composite expertise scores, ranging from −1.00 (novice) to 1.00 (expert). Emotional and cognitive responses were also measured via questionnaires.

Among 87 students completing the study, those in the AI-augmented personalized feedback group achieved significantly higher expertise scores compared with controls, including trial 5 (mean difference, 0.26; 95% CI, 0.09-0.43; P = .01) and in the realistic surgical task (mean difference, 0.20; 95% CI, 0.06-0.34; P = .02). This group also outperformed both other groups in critical safety-related measures such as bleeding and injury risk. Emotional and cognitive measures further distinguished the groups, underscoring the impact of feedback type on learner experience.

As the authors concluded, “Personalized expert instruction resulted in enhanced surgical performance and skill transfer compared with intelligent tutor instruction, highlighting the importance of human input and participation in artificial intelligence–based surgical training.”

Reference:
Giglio B, Albeloushi A, Alhaj AK, et al. Artificial intelligence-augmented human instruction and surgical simulation performance: a randomized clinical trial. JAMA Surg. Published online August 6, 2025. doi:10.1001/jamasurg.2025.2564