Microsoft’s Mustafa Suleyman says its AI outperforms doctors on hard cases as he chases ‘medical superintelligence’
News Source
•Thu, 22 Oct 2026 18:43:53 +0000
📰 What Happened
Microsoft says its AI system beat general doctors at solving hard medical cases. The results were published Thursday in the journal Science. The system is called MAI-DxO.
DxO runs several AI agents that act like a panel of doctors. They ask a patient questions, order tests, and argue about the diagnosis. On complex cases from the New England Journal of Medicine, it was right 83% of the time. General doctors were right 36% of the time.
One catch: cases this hard are usually handled by specialist teams. Those specialists were not part of the test. Microsoft AI chief Mustafa Suleyman called the result a step toward 'medical superintelligence.'
🔍 The Backstory
Microsoft is not alone in this claim. Google and OpenAI have also said their AI can match or beat doctors on some tests. The race to sell AI to hospitals is heating up fast.
The study's own data shows that stronger base AI models did most of the work. Microsoft's setup, which links agents together, added more gains on top. Microsoft tested it with models from OpenAI, Google, Anthropic, Meta, and DeepSeek.
Dominic King, who leads health at Microsoft AI, said the real problem is that demand for care is growing faster than the supply of doctors.
🎯 Why It Matters
If these results hold up in real clinics, AI could help patients get answers faster, especially in places short of doctors. But real patients are not test questions.
Microsoft says its AI system beat general doctors at solving hard medical cases. The results were published Thursday in the journal Science. The system is called MAI-DxO.
DxO runs several AI agents that act like a panel of doctors. They ask a patient questions, order tests, and argue about the diagnosis. On complex cases from the New England Journal of Medicine, it was right 83% of the time. General doctors were right 36% of the time.
One catch: cases this hard are usually handled by specialist teams. Those specialists were not part of the test. Microsoft AI chief Mustafa Suleyman called the result a step toward 'medical superintelligence.'
Microsoft is not alone in this claim. Google and OpenAI have also said their AI can match or beat doctors on some tests. The race to sell AI to hospitals is heating up fast.
The study's own data shows that stronger base AI models did most of the work. Microsoft's setup, which links agents together, added more gains on top. Microsoft tested it with models from OpenAI, Google, Anthropic, Meta, and DeepSeek.
Dominic King, who leads health at Microsoft AI, said the real problem is that demand for care is growing faster than the supply of doctors.
If these results hold up in real clinics, AI could help patients get answers faster, especially in places short of doctors. But real patients are not test questions.