Microsoft's MAI-DxO system, which has several AI agents act like a panel of doctors, correctly diagnosed difficult cases from the New England Journal of Medicine 83% of the time in a simulation, against 36% for generalist doctors working alone, according to a study in Science reported by TechCrunch. Specialists, who usually handle such cases, were not compared.
Microsoft AI chief Mustafa Suleyman called it a step towards medical superintelligence. The data suggests much of the gain comes from stronger underlying models, with the best results using OpenAI's GPT-5 and GPT-5.5. The orchestration helped most on complex cases and weaker models, and made little difference on emergency department cases.
Suleyman said Microsoft is working with the Mayo Clinic on a healthcare foundation model due next year. Google and OpenAI have made similar claims about their systems.
Source: TechCrunch
