Microsoft's MAI-DxO, a team of AI agents that question patients, order tests, and debate diagnoses, was correct on 83% of complex NEJM cases. Generalist physicians working alone got 36% right. The caveats: the cases were simulated, specialists weren't included, and most of the gain came from the underlying models. Mustafa Suleyman called the results a step toward "medical superintelligence," and Microsoft plans real-world testing with Beth Israel Deaconess. Would you trust an AI diagnosis over your doctor's, or only as a second opinion?
2