Microsoft has unveiled results from a major internal study claiming its medical AI system, dubbed the AI Diagnostic Orchestrator, outperformed experienced human doctors by a wide margin in diagnosing complex medical cases.
In a blog post released Monday, the company said the AI system correctly diagnosed nearly 86% of clinical cases tested—four times more accurately than human physicians working under the same constraints.
The tech giant’s announcement comes as artificial intelligence continues to push deeper into healthcare, raising both hopes for innovation and urgent questions about the future role of doctors in an AI-driven medical environment.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Inside the Study: AI vs. Doctors
The Microsoft-led study involved 304 real-world clinical case studies taken from the New England Journal of Medicine, a source widely respected for its complexity and diagnostic rigor. In the trial, the AI and 21 practicing physicians from the U.S. and U.K. were asked to diagnose the cases in stages—mirroring real-life practice—by ordering tests, asking follow-up questions, and narrowing down differential diagnoses.
Physicians participating in the study had between five and 20 years of clinical experience but were restricted from using tools they would typically rely on—like reference books, second opinions, or digital assistants.
While the doctors achieved an average diagnostic accuracy of just 20%, Microsoft’s AI system, when paired with OpenAI’s new o3 large language model, correctly diagnosed 85.5% of the cases. Microsoft tested the AI with other models too, including those from Meta, Anthropic, and Google, but found the strongest results in the OpenAI collaboration.
Cost, Accuracy, and Efficiency
Beyond accuracy, Microsoft also claimed that its AI solved cases more cost-effectively, raising the possibility of reducing waste in healthcare—a key concern in the U.S., where nearly 20% of GDP is spent on health, and up to 25% of that is believed to be unnecessary or wasteful.
In Microsoft’s view, AI could address both cost and diagnostic quality simultaneously.
“Our findings also suggest that AI can reduce unnecessary healthcare costs,” the company wrote, pointing to misdiagnosis, redundant testing, and administrative delays as areas ripe for disruption.
Mustafa Suleyman, head of Microsoft AI and cofounder of DeepMind, called the study a “big step toward medical superintelligence,” adding that the cases used were “some of the toughest and most diagnostically complex” challenges that doctors face.
Does This Mean AI Will Replace Doctors?
Despite the eye-catching headline numbers, Microsoft was careful to say that the goal is not to replace doctors, but to augment them.
“This technology represents a complement to doctors and other health professionals,” the company stated.
“While this technology is advancing rapidly, clinical roles are much broader than simply making a diagnosis. They need to navigate ambiguity and build trust with patients and their families in a way that AI isn’t set up to do,” Microsoft added.
Healthcare professionals appear to agree. Dr. Shravan Verma, CEO of a Singapore-based health tech startup, told Business Insider last month that AI tools may be well suited for handling the “first mile” of healthcare—triage, information gathering, or simple diagnosis—but emphasized that “AI can’t replicate physicians’ presence, empathy, and nuanced judgment in uncertain or complex conditions.”
However, others in tech are more bullish. Microsoft co-founder Bill Gates previously said on the People by WTF podcast in April that AI could eventually help solve the global shortage of doctors, arguing that AI systems could deliver “medical IQ” at scale.
A Broader Trend in AI and Healthcare
Microsoft is the latest in a growing list of tech giants racing to make inroads in the health sector. Google has developed its own medical AI tools and partnered with the Mayo Clinic. Amazon has expanded into telehealth. OpenAI, whose language model powers Microsoft’s system, has also acknowledged the medical space as a key future target.
The trend has not gone unnoticed by regulators and ethicists. With such powerful tools emerging, calls for guardrails around algorithmic bias, patient data privacy, medical liability, and the ethics of automation are also intensifying.
While Microsoft’s AI Diagnostic Orchestrator may mark a breakthrough in accuracy, the question remains how such tools will be integrated into real-world clinics, where human judgment, trust, and complexity still define much of the patient-doctor interaction.
Microsoft, for now, is signaling that it sees a future in which AI is not a replacement but a co-pilot for physicians—one capable of drastically improving diagnostic speed and quality if used responsibly.