Microsoft's AI diagnostic tool simulates a doctor's step-by-step questioning, with accuracy on complex cases about twice that of general practitioners

Cailian Press Financial News Agency

Open APP

Cailian Press Niu Zhanlin Editor-in-charge Gong Chuang 2026-10-23 21:55

① A Microsoft research team announced that an AI tool's diagnostic accuracy in complex case tests exceeded that of ordinary doctors by more than two times; ② The model uses a "diagnostic coordinator" mechanism to simulate a clinical expert team collaboratively analyzing cases and coordinate the entire diagnostic process; ③ This technology may also soon move toward commercialization, but Microsoft has not yet specified when it will be launched or in what form the service will be provided.

Cailian Press, October 23 (Editor Niu Zhanlin) For years, leaders in the field of artificial intelligence (AI) have regarded healthcare as one of the areas with the greatest promise for improving human life through AI. Now, this vision "is gradually becoming reality."

The related research was published in the journal Science, and the report shows that the model uses a mechanism called a "diagnostic coordinator" to decide which questions should be asked of the patient, which physical examinations and medical tests should be arranged, and to coordinate the entire diagnostic process. This mechanism simulates a team composed of multiple clinical experts, allowing different AI experts to collaboratively analyze cases, thereby organizing and advancing the diagnostic work.

In tests selected from hundreds of case reports in The New England Journal of Medicine, Microsoft's AI system achieved a diagnostic accuracy of about 83%; by comparison, general practitioners diagnosing with reference materials had an accuracy of about 36%, and doctors without reference materials had an accuracy of about 22%. When the test scenarios came from a large database of emergency admission cases, the tool also performed better than human doctors.

Under the coordination of the coordinator, the new AI model advances diagnosis step by step: first, it determines what medical history information about the patient needs to be learned, and which findings in the physical examination are significant; then, based on the information obtained, it decides which tests need to be arranged, analyzes the test results, and adds further tests as appropriate.

The head of the Microsoft research team said in an interview that during this process, the model continuously updates the current diagnostic hypotheses and adjusts the subsequent testing plan accordingly.

Microsoft said the system is powered by OpenAI's GPT-5. The researchers said the diagnostic coordinator also performed better than frontier AI models running independently. However, the peer-reviewed paper did not include newer models for comparison, such as GPT-6 Astra, Meta's Muse Wonder Full, Google's Gemini 4 Argon, and Anthropic's Opus 5.5.

Dominic King, vice president of Microsoft's AI healthcare business unit, said, "AI is already very good at this kind of test, and getting a perfect score is no longer difficult. But of course this is not the real-world situation. No patient comes to see me and directly hands over a clearly organized summary of their condition."

The researchers also found that compared with diagnostic testing plans formulated by human doctors, planning the diagnostic process by the coordinator costs less.

"What we measure is not only the system's diagnostic accuracy, but also whether it can prudently and carefully select test items and reasonably arrange diagnostic steps throughout the entire diagnostic process," said Harsha Nori, head of Microsoft's AI healthcare team.

Next, the research team will work with Beth Israel Deaconess Medical Center in Boston to test the tool's ability to triage patients in the emergency department waiting area.

This technology may also soon move toward commercialization, but Microsoft has not yet specified when it will be launched or in what form the service will be provided. Microsoft currently holds a considerable market share in the U.S. healthcare sector through digital tools, and healthcare is also one of the most popular application scenarios for its consumer-facing chatbot Copilot.

Global Market Intelligence

U.S. Stock Dynamics

Cailian Press Statement: The content of this article is for reference only and does not constitute investment advice. Investors operate at their own risk based on this.

Business Cooperation

Popular Unlocks

[Institutional Dragon-Tiger List Interpretation] Embodied intelligence + machine vision, has built a complete robot vision solution system and achieved breakthrough progress, jointly creating the Guangzhi embodied industrial solution with Yuejiang Technology, institutions made large net purchases of this company

① Embodied intelligence + machine vision, has built a complete robot vision solution system and achieved breakthrough progress, jointly creating the Guangzhi embodied industrial solution with Yuejiang Technology, institutions made large net purchases of this company; ② Embodied intelligence datasets + AI film and television + AI marketing + computing power, multiple embodied intelligence datasets completed data asset registration, the 3D articulated dataset was selected as a typical case of Beijing's high-quality datasets, this company received net purchases.

[Cailian Press Early Knowledge Weekly Review] Mentioned three times, 20cm limit-up the next day three times, cumulative increase this week exceeded 94%! The column efficiently decodes for you the logic behind the surge of this popular company

① The column's sub-modules jointly interpret, this popular company has both physical AI and robotics hotspots, mentioned three times, 20cm limit-up the next day three times, cumulative increase this week exceeded 94%; ② The column tracks relevant developments of overseas giants, sorts out key links in the robotics industry, and the related company gained 2 consecutive limit-ups after interpretation; ③ Data elements heated up this week, the column interpreted major policies, and the mentioned focus company rose more than 37% at its highest in 3 days.

① A table selects the high-quality information sorted out this week, with multiple popular companies in fields such as physical AI, robotics, and data elements on the list. The popular company's stock price nearly doubled this week; ② The robotics concept remained active, the column tracked domestic and international industry developments, and the two focus companies in the sector mentioned both hit limit-up; ③ The approval of the establishment of the National Data Group catalyzed the sector's heating up, the column interpreted the policy from an institutional perspective, and the related company mentioned hit limit-up.

① Tech themes attempted to stabilize and rebound, but the three major stock index futures still came under pressure and fell; ② Dan Bin's latest U.S. stock holdings: opened a position of 105,000 shares in SpaceX; ③ Goldman Sachs lowered Western Digital's target price; ④ AI commercialization accelerated, cloud communications company RingCentral rose about 15% premarket.

BofA: Market risk-aversion sentiment may continue until mid-November, U.S. stocks, the dollar, and U.S. Treasury yields may all fall together

Cailian Press Li Ying 2026-10-23 19:33

① BofA Chief Investment Strategist Michael Hartnett wrote in a report: "Until mid-November, we tactically maintain a risk-off stance"; ② He expects that U.S. Treasury yields, U.S. stocks, and the dollar will decline simultaneously next; ③ Investors are still continuously pouring into AI-related sectors, but in recent days have also begun to turn defensive.

Multiple vote banks turn into swing states! Trump again promises $150 million to increase fuel production

Cailian Press Ma Lan 2026-10-23 16:51

① U.S. President Trump on Thursday invoked the Defense Production Act and announced that the government will allocate $150 million to increase domestic gasoline and diesel production; ② The U.S. Department of Energy pointed out that this funding is expected to increase U.S. refinery output by millions of gallons per day; ③ As of now, U.S. gasoline prices remain above $4 per gallon, while diesel prices are about $5.4 per gallon.