Researchers at McGill University have demonstrated that combining artificial intelligence with expert human instruction delivers superior results compared to AI-only or traditional teaching methods in neurosurgical simulation training, with significant implications for advanced high-stakes technical education across multiple disciplines.
A groundbreaking study from The Neuro (Montreal Neurological Institute-Hospital) at McGill University has revealed that the most effective approach to training neurosurgeons combines artificial intelligence analytics with human expertise, challenging assumptions about AI replacing traditional instruction methods.
The research, published in JAMA Surgery on 6 August 2025, examined 87 medical students from four Quebec medical schools using virtual reality neurosurgical simulators. The findings suggest that whilst AI tutoring systems can provide valuable data-driven insights, they achieve optimal results when integrated with experienced human instruction rather than operating independently.
Comparative training methodology reveals hybrid approach superiority
The study employed a three-group experimental design to evaluate different training approaches. Students were divided into groups receiving AI-only verbal feedback, traditional expert instructor feedback, and a hybrid model where expert instructors utilised real-time AI performance data to inform their teaching.
The hybrid group demonstrated measurably superior performance across multiple metrics, including surgical skill acquisition speed, technique refinement, and critically, skill transfer capabilities. These students also showed significantly enhanced risk management abilities, particularly regarding bleeding control and tissue preservation – fundamental competencies in neurosurgical practice.
Technical integration enhances educational outcomes
The research builds upon previous work at the Neurosurgical Simulation and Artificial Intelligence Learning Centre, where AI tutoring systems had initially outperformed traditional human instruction. However, this earlier comparison involved human instructors working without access to AI-generated performance analytics.
“Our findings underscore the importance of human input in AI-driven surgical education,” explained lead author Bianca Giglio. “When expert instructors used AI performance data to deliver tailored, real-time feedback, trainees learned faster and transferred their skills more effectively.”
The virtual reality simulation environment enables comprehensive monitoring of student performance whilst providing immediate analytical feedback on surgical technique, decision-making, and procedural accuracy. This continuous assessment capability allows for unprecedented precision in identifying areas requiring improvement.
Implications extend beyond neurosurgical education
The study’s methodology and findings carry significant implications for technical training programmes across various high-stakes professions. The research demonstrates that AI systems excel at providing consistent, data-driven performance analysis, whilst human instructors contribute essential contextualisation, adaptability, and engagement that pure AI systems cannot replicate.
Dr Rolando Del Maestro, senior author and Director of the Centre, emphasised the collaborative nature of this educational approach: “AI is not replacing educators – it’s empowering them. By merging AI’s analytical power with the critical guidance of experienced instructors, we are moving closer to creating the ‘Intelligent Operating Room’ of the future capable of assessing and training learners whilst minimising errors during human surgical procedures.”
Future directions in AI-augmented medical training
The research suggests that optimal educational outcomes emerge from strategic integration of AI capabilities with human expertise rather than wholesale replacement of traditional teaching methods. This finding challenges prevalent assumptions about AI’s role in professional education and suggests a more nuanced approach to implementing artificial intelligence in training environments.
The study’s focus on skill transfer – the ability to apply learned techniques in new contexts – represents a crucial development in understanding how complex technical skills develop. This metric proves particularly relevant for neurosurgical training, where practitioners must adapt established techniques to unique patient presentations and varying clinical scenarios.
The research programme was supported by multiple funding bodies, including the Brain Tumour Foundation of Canada, the Royal College of Physicians and Surgeons of Canada, and several Quebec-based research foundations, reflecting the collaborative nature of advancing neurosurgical education through technological innovation.
Reference:
Giglio, B., Del Maestro, R., et al. (2025). Artificial intelligence augmented human instruction and surgical simulation performance. JAMA Surgery. Published online August 6, 2025. https://doi.org/10.1001/jamasurg.2025.2564
