{"slug": "a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36", "title": "A New Microsoft AI Diagnosed 83% Of Complex Medical Cases. Doctors Using References Got 36% Right.", "summary": "Microsoft researchers reported in Science that their MAI-DxO system, which coordinates multiple AI agents as virtual physicians, diagnosed 83% of 304 complex New England Journal of Medicine cases, versus 36% for generalist physicians given textbooks and online references and 22% without them. One configuration hit 83.2% accuracy at an estimated diagnostic cost of $5,431, while a cost-constrained version held 80.9% at $2,241, against $4,725 for physicians' modeled workups; on MIMIC-IV-ED emergency cases the system reached about 80% versus 47% for generalists. Microsoft is now preparing a prospective clinical study with Beth Israel Deaconess Medical Center in Boston, though the published results came from historical, text-based cases rather than live consultations.", "body_md": "# A New Microsoft AI Diagnosed 83% Of Complex Medical Cases. Doctors Using References Got 36% Right.\n\n## Microsoft is preparing a hospital study in Boston to assess how the technology performs alongside clinicians.\n\n[Microsoft](https://www.ibtimes.com/topic/microsoft) researchers have developed an [artificial intelligence](https://www.ibtimes.com/topic/ai) system that diagnosed 83% of complex medical cases in a controlled test, compared with 36% for general physicians using textbooks and online references.\n\nThe findings, published in the journal [Science](https://www.science.org/doi/10.1126/science.aeh7930), come from a system designed to investigate medical problems step by step rather than receive all the information needed for a diagnosis at once.\n\nCalled Microsoft AI Diagnostic Orchestrator, or MAI-DxO, the system coordinates multiple AI agents that act as a team of virtual physicians, assessing possible conditions and deciding which information is needed before reaching a diagnosis.\n\nResearchers evaluated the technology using 304 complicated cases published in *The New England Journal of Medicine*. Participating generalist physicians correctly diagnosed 22% without outside reference materials and 36% when they could consult them, according to [Microsoft AI's announcement](https://microsoft.ai/news/building-medical-superintelligence/).\n\nTraditional medical AI tests often use multiple-choice questions that provide all the relevant clinical information upfront. Microsoft's researchers instead developed the Sequential Diagnosis Benchmark, or SDBench, to assess how AI investigates an unfamiliar patient.\n\nStarting with a short clinical description, MAI-DxO can request medical history, physical examination findings and laboratory or imaging tests. Its virtual specialists propose diagnoses, challenge one another's assessments and determine whether further investigation is necessary. The benchmark measures diagnostic accuracy, the information requested and the estimated cost of testing.\n\nResearchers also evaluated the system using emergency department cases from MIMIC-IV-ED, a database containing deidentified hospital records. MAI-DxO achieved approximately 80% accuracy, compared with 47% for generalist physicians under the study conditions.\n\nThe results extended beyond diagnostic performance. Microsoft examined whether the system could reach accurate conclusions while limiting unnecessary medical investigations. Each requested test was assigned an estimated cost, allowing researchers to compare different approaches.\n\nOne configuration achieved 83.2% accuracy on complex cases at an estimated diagnostic cost of $5,431. When stronger cost constraints were introduced, accuracy remained at 80.9%, while expenses fell to $2,241.\n\nThe physicians' comparative workups averaged $4,725. These were modeled testing costs, not savings demonstrated in hospitals. Microsoft tested its orchestration method with language models from several developers, including OpenAI, Google, Anthropic, Meta and DeepSeek.\n\nSome models performed well independently. For example, GPT-5 achieved approximately 79% accuracy on the complex cases without the orchestration system, compared with just over 83% when used with MAI-DxO.\n\nMicrosoft said subsequent evaluations involving newer models, including GPT-5.5, also showed improvements when the orchestration method was applied.\n\nHowever, the research was conducted using historical, text-based cases rather than live clinical consultations. The physician comparison involved generalists working individually, not specialist teams.\n\nMicrosoft is now working with Beth Israel Deaconess Medical Center in Boston to test the system in clinical settings.\n\nThe planned prospective study will involve patients interacting directly with the technology while receiving conventional medical care. Researchers will compare its assessments with physicians' independent evaluations to identify where the system performs well and where it encounters difficulties.\n\n© Copyright IBTimes 2026. All rights reserved.", "url": "https://wpnews.pro/news/a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36", "canonical_source": "https://www.ibtimes.com/new-microsoft-ai-diagnosed-83-complex-medical-cases-doctors-using-references-got-36-right-3808909", "published_at": "2026-10-23 15:11:44+00:00", "updated_at": "2026-10-23 16:12:58.444323+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-agents", "ai-research", "large-language-models", "ai-safety"], "entities": ["Microsoft", "MAI-DxO", "Science", "Beth Israel Deaconess Medical Center", "OpenAI", "GPT-5", "Google", "Anthropic"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36", "markdown": "https://wpnews.pro/news/a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36.md", "text": "https://wpnews.pro/news/a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36.txt", "jsonld": "https://wpnews.pro/news/a-new-microsoft-ai-diagnosed-83-of-complex-medical-cases-doctors-using-got-36.jsonld"}}