Home / Science / AI Diagnoses Disease 4x More Accurately Than Doctors
AI Diagnoses Disease 4x More Accurately Than Doctors
22 Oct
Summary
- AI tool diagnoses conditions up to four times more accurately.
- System achieved 83% accuracy on New England Journal cases.
- Microsoft plans hospital testing within months.
Microsoft AI CEO Mustafa Suleyman announced a significant advancement in AI-powered healthcare diagnostics, with a new tool demonstrating diagnostic accuracy up to four times greater than human doctors.
Published in Science, the AI system, called the Diagnostic Orchestrator, was tested against human physicians using medical case reports and emergency department data. On cases from The New England Journal of Medicine, the AI achieved an 83% accuracy rate, compared to 36% for doctors using references and 22% for those without.
Further tests using data from Beth Israel Deaconess Medical Center showed the AI scoring 80% accuracy, well above the 47% average for human physicians. The system, powered by OpenAI's GPT-5, also tended to reach diagnoses at a lower estimated cost.
Microsoft acknowledges that clinical validation in real-world settings is crucial before widespread deployment. The company is planning to begin pilot testing the system in hospitals with Beth Israel Deaconess Medical Center within the next couple of months.
This development is part of Microsoft's larger goal to create 'medical superintelligence' that surpasses human-led healthcare systems. While AI aims to assist and augment, company leaders emphasize that human clinicians will always retain a central role in patient care.
