In a groundbreaking advance set to redefine surgical education, researchers at Mount Sinai have successfully demonstrated that complex surgical procedures can be taught to trainees using an autonomous system powered by artificial intelligence (AI) and extended reality (XR) technology—completely eliminating the need for an in-person instructor. This novel approach, which leverages deep learning algorithms in concert with a custom-designed XR headset, was tested on 17 surgical trainees, all of whom achieved surgical success on a highly intricate simulated nephrectomy procedure. The implications of this study extend far beyond resident training, opening doors to a future where autonomous, AI-driven education can play a pivotal role across various medical disciplines.
Traditionally, surgical training necessitates the presence of experienced proctors who guide trainees through hands-on procedures in an operating room setting. This model, while effective, has inherent limitations including the scarcity of qualified instructors, inconsistent skill transmission, high training costs, and the logistical hurdles of coordinating in-person teaching. Mount Sinai’s research confronts these issues head-on by introducing an innovative AI-augmented instructional system that integrates seamlessly with XR visualization tools, enabling trainees to receive tailored, real-time guidance without direct human supervision. The study’s encouraging results indicate a significant leap forward in democratizing surgical education.
Central to the study is the development of an AI algorithm linked to an extended-reality headset that continuously monitors a trainee’s performance during a simulated partial nephrectomy—a procedure involving the careful removal of a cancerous portion of a kidney along with vascular clamping of the renal artery. The kidney model used in training was created through sophisticated 3D printing technology based on anonymized patient CT scans, assembled with water-based polymers to mimic real tissue consistency and anatomical features. This material authenticity allowed for highly realistic practice conditions, enabling the AI system to accurately evaluate and provide corrective feedback on the trainee’s technique, positioning, and decision-making in real-time.
The AI system, referred to as ESIST (educational system for instructionless surgical training), employs deep learning methodologies to interpret data captured from a first-person perspective camera embedded in the XR headset. This camera tracks hand movements, tool positioning, and procedural progression. Using this data, the AI assesses whether the trainee is correctly following each step of the protocol and projects corrective prompts directly into the visual field of the headset. This unique fusion of AI assessment and visual augmentation creates an immersive, hands-free learning experience that finely balances instruction with active practice, thus facilitating highly effective skill acquisition and retention.
Dr. Nelson Stone, Clinical Professor of Urology, Radiation Oncology, and Oncological Sciences at the Icahn School of Medicine at Mount Sinai and corresponding author of the study, highlights the remarkable accuracy of the system, stating that it achieved a 99.9 percent precision rate in guiding the critical steps of the procedure. This milestone illustrates not only the technological sophistication of AI-driven surgical education but also its immense potential to standardize training outcomes, minimizing variability that often arises from human instruction and personal teaching styles. Such standardization could be instrumental in ensuring uniformly high-quality surgical care across diverse healthcare settings.
One of the underpinning motivations behind this research is the pressing shortage of qualified surgical instructors and supervisors, compounded by increasing demand for new training on advanced surgical devices and techniques. Mount Sinai’s autonomous educational model presents a scalable solution capable of alleviating bottlenecks caused by instructor availability and time constraints of attending physicians. By delivering expert-level guidance autonomously, trainees can engage in repetitive practice without the immediate presence of a mentor, accelerating proficiency and freeing instructors to focus on higher-level supervisory roles.
Moreover, the use of autonomous AI training outside of the operating room holds profound implications for patient safety. By enabling surgeons to refine their skills extensively in simulated environments, the risk of intraoperative errors can be substantially diminished. As trainees gain confidence and precision through iterative practice with AI feedback, the likelihood of complications during real surgeries may decrease significantly—ultimately translating into better patient outcomes and enhanced healthcare quality.
The success of Mount Sinai’s system, however, is not limited solely to its current capabilities. The research team aims to advance this technology further by developing more complex synthetic cadaver models that replicate entire surgical procedures instead of isolated components. This evolution will allow trainees to practice comprehensive operations from start to finish under AI-guided supervision, further bridging the gap between simulation and live surgery. As the system matures, it also holds promise for adapting to diverse surgical specialties, potentially revolutionizing medical education on a global scale.
Following the initial training phase, participant feedback underscored the educational impact of the AI-XR system. Remarkably, all trainees agreed that the program offered significant instructional value, confirming that immersive, real-time AI guidance can effectively supplement—and in some cases, substitute—the traditional apprenticeship model. This unanimous positive reception underscores the system’s usability and appeals to the next generation of surgeons, who are increasingly comfortable with technology-driven learning.
The integration of cutting-edge AI algorithms with extended reality represents a convergence of several advanced technological domains including computer vision, machine learning, medical imaging, and virtual simulation. By bringing these fields together to address a critical bottleneck in surgical education, Mount Sinai’s team has ignited a new pathway towards autonomous and highly scalable medical training solutions. This interdisciplinary approach exemplifies how emerging technologies can be harnessed for practical applications with profound societal benefits.
Financially, the innovation has substantial ramifications. By reducing reliance on proctors and enabling remote or unsupervised training, training programs could drastically cut expenditures related to personnel costs, operating room availability, and associated administrative overhead. These savings create an avenue for investing more resources into the refinement of training models, acquisition of advanced simulation hardware, and expansion of educational accessibility worldwide—particularly in regions where expert surgical mentorship is scarce.
Despite the promising outcomes, the researchers emphasize the need for continued studies involving larger cohorts and diverse procedural training to validate and optimize the system’s efficacy broadly. Regulatory, ethical, and practical considerations must also be addressed as this autonomous technology moves closer to integration into accredited surgical education curricula. Nonetheless, Mount Sinai’s initial demonstration stands as a compelling proof-of-concept that autonomous AI-guided education is not merely feasible but capable of transformative impact.
In conclusion, the Mount Sinai researchers have charted an exhilarating course toward the future of surgical training, where artificial intelligence and immersive reality technologies collaborate to reshape how skills are taught and mastered. Their pioneering work offers a vision of medical education that is more efficient, standardized, accessible, and patient-centric. As AI systems evolve from adjunctive tools to primary instructors within highly simulated environments, the next generation of surgeons may arise more skilled, confident, and prepared for the challenges of modern medicine than ever before.
Subject of Research: Not applicable
Article Title: Autonomous Educational System for Surgical Training Utilizing Deep Learning Combined with Extended Reality
News Publication Date: August 6, 2025
Web References: https://www.liebertpub.com/doi/10.1177/29941520251361898
References: Jonathan J. Stone, Nelson N. Stone, Steven H. Griffith, Kyle Zeller, Michael P. Wilson, Journal of Medical Extended Reality, 23-Jul-2025
Keywords: Machine learning, Adaptive systems
