Blog

Microsoft has revealed details of an artificial intelligence system that performs better than human doctors at complex health diagnoses, creating a “path to medical superintelligence”.

The company’s AI unit, which is led by the British tech pioneer Mustafa Suleyman, has developed a system that imitates a panel of expert physicians tackling “diagnostically complex and intellectually demanding” cases.

Microsoft said that when paired with OpenAI’s advanced o3 AI model, its approach “solved” more than eight of 10 case studies specially chosen for the diagnostic challenge. When those case studies were tried on practising physicians – who had no access to colleagues, textbooks or chatbots – the accuracy rate was two out of 10.

Microsoft said it was also a cheaper option than using human doctors because it was more efficient at ordering tests.

Despite highlighting the potential cost savings from its research, Microsoft played down the job implications, saying it believed AI would complement doctors’ roles rather than replace them.

“Their clinical roles are much broader than simply making a diagnosis. They need to navigate ambiguity and build trust with patients and their families in a way that AI isn’t set up to do,” the company wrote in a blogpost announcing the research, which is being submitted for peer review.

However, using the slogan “path to medical superintelligence” raises the prospect of radical change in the healthcare market. While artificial general intelligence (AGI) refers to systems that match human cognitive abilities at any given task, superintelligence is an equally theoretical term referring to a system that exceeds human intellectual performance across the board.

Suleyman, the chief executive of Microsoft AI, told the Guardian the system would be operating perfectly within the next decade.

“It’s pretty clear that we are on a path to these systems getting almost error-free in the next 5-10 years. It will be a massive weight off the shoulders of all health systems around the world,” he said.

Explaining the rationale behind the research, Microsoft raised doubt over AI’s ability to score exceptionally well in the United States Medical Licensing Examination, a key test for obtaining a medical licence in the US. It said the multiple-choice tests favoured memorising answers over deep understanding of a subject, which could help “overstate” the competence of an AI model.

Microsoft said it was developing a system that, like a real-world clinician, takes step-by-step measures – such as asking specific questions and requesting diagnostic tests – to arrive at a final diagnosis. For instance, a patient with symptoms of a cough and fever may require blood tests and a chest X-ray before the doctor arrives at a diagnosis of pneumonia.

The new Microsoft approach uses complex case studies from the New England Journal of Medicine (NEJM).

Suleyman’s team transformed more than 300 of these studies into “interactive case challenges” that it used to test its approach. Microsoft’s approach used existing AI models, including those produced by ChatGPT’s developer, OpenAI, Mark Zuckerberg’s Meta, Anthropic, Elon Musk’s Grok and Google’s Gemini.

Microsoft then used a bespoke, agent-like AI system called a “diagnostic orchestrator” to work with a given model on what tests to order and what the diagnosis might be. The orchestrator in effect imitates a panel of physicians, which then comes up with the diagnosis.

Microsoft said that when paired with OpenAI’s advanced o3 model, it “solved” more than eight of 10 NEJM case studies – compared with a two out of 10 success rate for human doctors.

Microsoft said its approach was able to wield a “breadth and depth of expertise” that went beyond individual physicians because it could span multiple medical disciplines.

It added: “Scaling this level of reasoning – and beyond – has the potential to reshape healthcare. AI could empower patients to self-manage routine aspects of care and equip clinicians with advanced decision support for complex cases.”

Microsoft acknowledged its work is not ready for clinical use. Further testing is needed on its “orchestrator” to assess its performance on more common symptoms, for instance.

By Greg Robb

Treasury Secretary says Fed seems ‘frozen at the wheel’

President Donald Trump has continued to put pressure on Federal Reserve Chair Jerome Powell to dramatically lower interest rates.

At a Monday press briefing, White House Press Secretary Karoline Leavitt held up a handwritten note to the Fed leader from the president.

The letter suggested that Trump thinks the Fed’s benchmark rate should be between close to 0.5%, down from its current range of 4.25% to 4.5%.

The note read: “Jerome, You are, as usual, ‘too late.’ You have cost the USA a fortune – and continue to do so – You should lower the rate – by a lot. Hundreds of billions of dollars being lost. No Inflation”

Powell has been pushing back on rate cuts all year, saying he wants to wait to learn more about “the likely course of the economy” from the president’s tariff policy.

The Fed will meet again in July to consider a rate cut. The market is pricing in only about a 20% chance of a cut at that meeting, after a key inflation measure released last week was slightly hotter than expected.

Traders are now pricing in three quarter-point cuts at the Fed’s final three meetings of the year, starting in September. Analysts at JPMorgan say the reasons behind the cuts may not be supportive for stocks.

Earlier Monday, other top White House officials also kept up the criticism of Powell.

Treasury Secretary Scott Bessent said in a Bloomberg interview that Fed officials “seem a little frozen at the wheel.”

Bessent said the Fed was too worried about making another mistake like its failure to see high inflation brewing in 2022.

Powell’s term as Fed chair ends next May. In the interview, Bessent said the White House will be working “over coming weeks and months” on choosing Powell’s successor.

Additionally, Fed Governor Adriana Kugler’s term at the central bank is up at the end of January. Bessent said the White House might use the vacancy created by Kugler’s departure to put Powell’s replacement on the Fed’s seven-member board of governors.

Bessent is considered to be on the short list of possible replacements for Powell.

Asked about this possibility, Bessent said: “I will do what the president wants, but I think I have the best job in D.C.”

-Greg Robb

This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.

  (END) Dow Jones Newswires

  06-30-25 1609ET
Copyright (c) 2025 Dow Jones & Company, Inc.

As treatment approaches in multiple myeloma continue to advance, new areas of controversy arise, leading to undefined expectations within clinical practice, according to Ajai Chari, MD.

Accordingly, experts gathered at the inaugural Bridging the Gaps in Leukemia, Lymphoma, and Multiple Myeloma Conference to identify current controversies or challenges in the field and come to a consensus about the optimal approaches to myeloma management. An overview of these findings was published in a manuscript after the meeting.

“There’s a joke in that if you ask myeloma doctors for an opinion, you’ll get 6 answers. This [meeting] shows that there’s a lot of complexity and evolving data [in the multiple myeloma space],” Ajai Chari, MD, explained in an interview with OncLive®.

During the interview, Chari detailed the importance of publishing this manuscript; the limitations to current risk stratification approaches in smoldering multiple myeloma; ongoing challenges regarding determination of the optimal timing for treatment initiation; and what the potential FDA approval of daratumumab (Darzalex) could mean for this patient population.

Chari is director of the Multiple Myeloma Program and a professor of medicine in the School of Medicine at the University of California, San Francisco.

OncLive: What is the importance of publishing this manuscript from the Bridging the Gaps in Leukemia, Lymphoma, and Multiple Myeloma?

Chari: The purpose of this manuscript is to put a bunch of experts in the room and see how people think, where we agree, and where the areas of controversy are that we need to work on. It’s really for people who may not be doing this day in and day out, [in which we] take a little look under the hood, [so to speak].

How is risk stratification approached for smoldering multiple myeloma? Are there any limitations to current models?

Conventionally, risk stratification for smoldering has been based on static models. For example, how much M-spike [is present]? What’s the light chain? What’s the bone marrow burden? What are the immunoglobulin levels? The limitation there is that the date of diagnosis of smoldering myeloma is typically the date of the marrow, but somebody could have been smoldering prior to the marrow for anywhere from a month or 2 years. When you try to figure out who’s going to progress in the short term, that’s a big confounding variable. Although we need static models, because they’re convenient, we need more dynamic models, which tell us how the tumor is evolving. Are the protein levels going up? Is the hemoglobin drifting down? Are the bone lesions evolving? Those are where the field needs to go. Also, incorporating genetics, immune markers, and genomic markers. We haven’t gotten there yet, and that’s the struggle of serious stratification.

What are some key considerations when deciding on an optimal time to initiate therapy and treatment strategies in smoldering multiple myeloma?

The struggle with treating patients with smoldering myeloma is that they are essentially healthy people. They don’t have active myeloma. [The important questions are:] how do you not over-treat these patients with potentially toxic therapies that could cause adverse effects to help the individuals? Also, how do we not under-treat these patients? If we have amazing therapies that can help prevent symptoms and the onset of myeloma, shouldn’t we be doing that? That’s the tightrope that we have to walk in figuring out what the right time to initiate is. That’s where the risk stratification can help us make that decision.

What would the FDA approval of daratumumab mean for this space?

[Data from] the [phase 3] AQUILA study [NCT03301220] published in The New England Journal of Medicine showed, in a randomized fashion, that observation vs treatment with daratumumab…led to improved response, including progression-free survival and, surprisingly, overall survival. We just heard that the European Medical Association voted favorably on it; if that’s FDA approved, we could give patients with smoldering myeloma a relatively well-tolerated therapy, [particularly those who] we are concerned about progressing.

[Nevertheless], I’m most excited [to potentially use daratumumab plus lenalidomide [Revlimid], bortezomib [Velcade], and dexamethasone [Dara-VRd] for older patients who have smoldering myeloma and may have a high risk of progression. [These patients likely] would not be excited about having to receive a multidrug therapy in the future. [Therefore], it’s very reasonable to consider this

[Daratumumab] is a monoclonal antibody to help prevent the development of myeloma and potentially confer those clinical benefits. The challenge is that the study included 3 years of therapy, and that’s a [relatively] long period. Where I struggled more is with the younger patients; where they are now, they could easily tolerate multidrug therapy. However, do we need to commit everybody to 3 years of preventative therapy? This is another tightrope we have to walk.

Reference

Chari A, Bal S, Ailawadhi S, et al. Expert Perspectives on Current Challenges and Emerging Approaches for Multiple Myeloma: Narrative Review of an Inaugural Bridging the Gaps in Leukemia, Lymphoma, and Multiple Myeloma. Clin Lymphoma Myeloma Leuk. Published online March 11, 2025. doi:10.1016/j.clml.2025.03.008