The COVID-19 epidemic has presented both opportunities and challenges for medical training. In many fields, remote learning technology is becoming more important. A new study shows that AI tutoring systems can outperform human instructors in remote settings.
The Neurosurgical Simulation and Artificial intelligence Learning Centre at The Neuro (Montreal Neurological Hospital) recruited seventy medical students for virtual brain tumour removals using a neurosurgical simulator. The students were randomly assigned to receive instruction or feedback from either an AI tutor, or a remote expert instructor. A third control group was not provided with any instruction.
An AI-powered tutor named the Virtual Operative Assistant, (VOA), used a machine-learning algorithm to teach safe surgical techniques and provided personalized feedback. A deep learning Intelligent Continuous Expertise Monitoring System and a panel assessing student performance also assessed student performances.
Remote instructors viewed a live stream of the surgical simulations in the other group and gave feedback based upon the student’s performance.
Researchers found that VOA students learned surgical skills 2.6x faster and performed 36% better than those who received feedback and instruction from remote instructors. Researchers expected VOA students to feel more stress and negative emotion. However, there was no significant difference in the results.
Both during and after brain surgery, surgical skill is crucial for patient outcomes. VOA could be a good way to increase neurosurgeon’s performance, improve patient safety, and reduce the burden for human instructors.
The VOA is an artificially intelligent tutor that could be a valuable tool for training the next generation of neurosurgeons. The VOA helped improve expertise and foster a learning environment. In-depth studies are ongoing to determine how AI-powered intelligent tutors and in-person instructors can be combined to improve neurosurgical skills.
Dr. Rolando Del Maestro, study’s senior author
Ali Fazlollahi is the study’s first author. He stated that intelligent tutoring systems can use a variety simulation platforms to offer almost unlimited opportunities for repetitive practice without the limitations imposed by supervision. “We can be better prepared to meet ever-changing future challenges by continuing research, increasing development, and disseminating intelligent tutoring systems.”
