Boston:Microsoft has announced that its artificial intelligence system has achieved more than double the accuracy of doctors in diagnosing complex medical conditions during simulated tests, as per research published in the journal Science.

According to Anadolu Agency – English, Microsoft's AI Diagnostic Orchestrator (MAI-DxO) attained an 83% diagnostic accuracy rate in complex medical cases, significantly surpassing the 36% accuracy rate of physicians who relied on textbooks and online resources. When physicians worked without external reference materials, their accuracy dropped to 22%.

The evaluation involved 304 complex cases sourced from The New England Journal of Medicine. The AI system utilized multiple agents working collaboratively to assess symptoms, choose diagnostic tests, and draw conclusions. Additionally, the AI system achieved 80% accuracy in more typical emergency department scenarios compared to 47% for doctors.

Microsoft indicated that the MAI-DxO system could reduce diagnostic testing costs by opting for fewer or less expensive procedures without compromising accuracy. In a cost-optimized setup, the system maintained an 81% accuracy rate on complex cases at less than half the cost of existing comparative systems.

Describing these findings as a step toward achieving "medical superintelligence," Microsoft envisions AI systems capable of exceeding the capabilities of top clinical teams in medical reasoning and care coordination. However, it noted that these results were based on simulated cases and not real-world clinical practice, with the physician comparison not including specialist teams.

Microsoft emphasized the necessity for further evaluation in clinical environments to determine the system's performance in actual patient care. The company plans a prospective study with Beth Israel Deaconess Medical Center in Boston to compare the AI system's diagnostic assessments with those made by physicians in hospital settings.