Microsoft has announced a groundbreaking achievement in healthcare with its AI system, the Microsoft AI Diagnostic Orchestrator (MAI-DxO), which has accurately diagnosed the majority of the world’s toughest medical cases. According to research published on June 30, 2025, MAI-DxO correctly solved 85.5% of 304 challenging cases featured in the New England Journal of Medicine—one of the highest accuracy rates ever recorded in the medical AI field.
The study compared MAI-DxO’s performance to that of 21 experienced physicians from the US and UK, who achieved an average accuracy rate of just 20%. Impressively, MAI-DxO did not only outperform them in accuracy—it also reached its results at a lower average diagnostic cost by ordering fewer unnecessary tests. Unlike static diagnosis tools, MAI-DxO simulates a virtual panel of five AI personas, acting as if it were multiple doctors debating, questioning, ordering specific tests, and challenging diagnostic assumptions. This ‘chain-of-debate’ method helps the AI reach greater diagnostic precision and avoids needless testing and expenses.
With US healthcare spending nearing 20% of GDP and up to 25% believed to be wasted on low-value tests, MAI-DxO’s efficiency and accuracy could substantially reduce costs and improve patient outcomes. Although MAI-DxO is not yet approved for clinical use, Microsoft emphasizes the transformative potential of this technology after it undergoes rigorous safety and real-world testing.
Microsoft’s clinical AI experts highlight that this system won’t replace doctors but will complement their expertise—automating routine decisions and enabling professionals to tackle highly complex cases with advanced support. The AI’s ability to enable more cost-effective, precise diagnosis may also empower patients to better manage their own care.
While these results are promising, Microsoft stresses that robust validation in real clinical environments and regulatory approval will be necessary before MAI-DxO is used in everyday practice. The company is partnering with leading health organizations to ensure AI is safely and effectively adopted, aiming to revolutionize healthcare delivery with human-machine collaboration.