We are pioneering the integration of AI and surgery to deliver safer, more precise, and more personalized care.
We leverage artificial intelligence and surgical data science at scale to transform surgical care, accelerate research and innovation, and improve patient outcomes.
At Mount Sinai, we are defining the Surgeon of the Future. By utilizing artificial intelligence (AI) to capture the data in the operating room from every action, instrument, and signal, surgeons can drive the creation of novel surgical AI tools. These tools enable surgeons to identify patterns unrecognizable with traditional vision and intervene beyond the reach of present-day dexterity. This empowerment translates to safer, more effective, and individualized surgical care for our patients.
A secure framework for learning from thousands of surgical videos to enable rigorous AI research and discovery.
AI learns by leveraging vast quantities of data. We are organizing these complicated data streams into an Electronic Surgical Record to enable efficient dialogue between AI systems, thus allowing for future technological breakthroughs.
Researching computer vision and edge computing tools in the operating room.
When AI tools are used well, they can enhance how surgeons see, think, and act in the operating room. Our surgeons are teaching the AI context-awareness in the operating room to build tools for surgical phase recognition, navigation assistance, and intraoperative decision support.
Utilizing AI to transform the way we diagnose and communicate with our patients.
We have already published research that is translating into patient benefit. Our ongoing work involves multimodal prediction models for cancer, agentic large language models for frailty assessment, and multi-agent tumor board simulation.
The backbone for the Program for Surgical Intelligence is a secure, standardized repository of surgical video and intraoperative data. Working in partnership with Mount Sinai's Windreich Department of Artificial Intelligence and Human Health, we leverage Mount Sinai's vast data science resources integrated with the Minerva supercomputer, one of the largest academic AI compute infrastructures in the country.
Responsible AI requires more than innovation — it requires trust. We are committed to:
By placing ethics at the core, we ensure that surgical AI adoption is safe, equitable, and patient-centered.
No US law defines who owns surgical data, leaving AI vendors free to commercialize OR video without patient consent; this JAMA Surgery piece proposes a custodianship framework and four federal policy actions to close that gap.
A hierarchical AI system performed key steps of a laparoscopic cholecystectomy fully autonomously with 100% success across 8 ex vivo gallbladders, marking a credible milestone on the road to clinically deployable surgical robots.
Built over seven years with global surgeon-researcher-industry collaboration, this framework defines a four-level hierarchy (phase, step, task, subtask) for annotating robotic surgical video and has already been applied to over 15,000 hours across 13,465 procedures.
A systematic review of all 49 FDA-cleared surgical robots from 2015 to 2023 found that 86% remain fully surgeon-controlled and only 2 are officially recognized as ML-enabled, exposing how far regulatory frameworks lag behind the technology.
An offline GPU edge device running a pre-trained neural network achieved 68.7% accuracy identifying surgical phases in real time during robotic inguinal hernia repair, representing the first demonstration of live computer vision feedback directly in the operating room.
AI trained on videos from 153 surgeons across 37 countries can identify safe and dangerous dissection zones during gallbladder surgery in real time with greater than 94% accuracy, offering a potential safeguard against bile duct injury.
By cataloguing over 34,000 robotic surgical gestures, researchers demonstrated that specific instrument-tissue interactions during nerve-sparing prostatectomy independently predict whether a patient will recover erectile function one year after surgery.
This landmark review traces computer vision from proof-of-concept research to clinical translation in minimally invasive surgery and maps the data, regulatory, and ethical barriers that still stand between the lab and the OR.
A view of the clinical application of AI at Mount Sinai from the vantage point of our leaders.
In a landmark validation study, Mayo Clinic's AI model identified pancreatic cancer on routine CT scans up to three years before clinical diagnosis by detecting subtle tissue changes invisible to the human eye.
Not yet — and probably not for quite some time. Here is the measured perspective of Dr. Daniel Herron, the Chief of the Garlock Division of General Surgery at Mount Sinai, on exciting developments out of Johns Hopkins.
A Reuters investigation found that FDA-authorized AI surgical devices have generated adverse event reports, raising urgent questions about how surgical AI is tested and overseen before reaching patients. PSI is keeping this top-of-mind as it seeks to develop responsible AI.
Congratulations to Dr. Lee for receiving a prestigious career development award from the NIH for her AI in Endocrine Surgery work!
AI agents acting as a collaborative virtual lab designed novel COVID-19 vaccine candidates in days, pointing toward a future where AI dramatically compresses the timeline from hypothesis to discovery.
Dr. Marin is a vascular surgeon, Chair of the Department of Surgery, and Surgeon-in-Chief at the Icahn School of Medicine at Mount Sinai. He pioneered the field of endovascular surgery in the 1990s and has been shepherding surgical breakthroughs to real world clinical impact ever since.
View Profile →
Dr. Lee is an NIH-funded endocrine surgeon-scientist exploring how AI can transform thyroid cancer management. Her research applies novel multimodal machine learning methods to refine preoperative risk prediction and guide personalized surveillance strategies.
View Profile →
Dr. Oland is a minimally invasive surgeon with a background in engineering and medical innovation, and a special interest in computer vision, a form of AI that teaches computers how to understand operative video. Daily he is taking care of patients or in the operating room, actively using and evaluating the tools that are inspiring the technological breakthroughs of tomorrow.
View Profile →
Dr. Nadkarni is System Chair of the Windreich Department of Artificial Intelligence and Human Health, and Director of the Hasso Plattner Institute for Digital Health. He is also the Chief AI Officer of the Mount Sinai Health System and therefore leads the system's clinical AI integration efforts, with a particular focus on translating data science and machine learning into scalable solutions that enhance patient care.
View Profile →
Dr. Glicksberg directs the Digital Discovery Program as part of the Hasso Plattner Institute for Digital Health and the Windreich Department of AI and Human Health, and is the Founding Director of the Center for AI in Children's Health. His work involves the analysis and integration of diverse multimodal data for applications ranging from predictive modeling to drug discovery.
View Profile →
Dr. Sawant works clinically in the Department of Medicine and as a clinical informaticist in the Windreich Department of AI and Human Health. Drawing on his frontline clinical experience, he ensures that data generated during patient care is organized, interoperable, and primed to power novel clinical models on the backend.
View Profile →
Steven works as a clinical informaticist at Mount Sinai and is responsible for creating machine learning pipelines and extracting electronic health record data used to conduct novel surgical research.
Join us in shaping the future of surgical AI. We welcome surgeons, trainees, data scientists, engineers, industry and academic collaborators, and patient advocacy groups.