Microsoft says its new health AI beat doctors in accurate diagnoses by a mile
The AI system also solved cases "more cost-effectively" than its human counterparts, Microsoft said.
The study comes as AI's growing role in healthcare raises questions about its place in medicine.
Microsoft said its medical AI system diagnosed cases more accurately than human doctors by a wide margin.
In a blog post published on Monday, the tech giant said its AI system, the Microsoft AI Diagnostic Orchestrator, diagnosed cases four times as accurately as a group of experienced physicians in a test.
Microsoft's study comes as AI tools rapidly make their way into hospitals and clinics, raising questions about how much of medicine can or should be automated and what role doctors will play as diagnostic AI systems get more capable.
The experiment involved 304 case studies sourced from the New England Journal of Medicine. Both the AI and physicians had to solve these cases step by step, just like they would in a real clinic: ordering tests, asking questions, and narrowing down possibilities.
The AI system was paired with large language models from tech companies like OpenAI, Meta, Anthropic, and Google. When coupled with OpenAI's o3, the AI diagnostic system correctly solved 85.5% of the cases, Microsoft said.
By contrast, 21 practicing physicians from the US and UK — each with five to 20 years of experience — averaged 20% accuracy across the completed cases, the company added. In the study, the doctors did not have access to resources they might typically tap for diagnostics, including coworkers, books, and AI.
The AI system also solved cases "more cost-effectively" than its human counterparts, Microsoft said.
"Our findings also suggest that AI reduce unnecessary healthcare costs. US health spending is nearing 20% of US GDP, with up to 25% of that estimated to be wasted," it added.
"We're taking a big step towards medical superintelligence," said Mustafa Suleyman, the CEO of Microsoft's AI division, in a post on X.
He added that the cases used in the study are "some of the toughest and most diagnostically complex" a physician can face.
Suleyman previously led AI efforts at Google.
Microsoft did not respond to a request for comment from Business Insider.
Microsoft said in the blog post that AI "represents a complement to doctors and other health professionals."
"While this technology is advancing rapidly, 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," Microsoft said.
"Clinical roles will, we believe, evolve with AI," it added.
Tech leaders like Microsoft cofounder Bill Gates have said that AI could help solve the long-standing shortage of doctors.
"AI will come in and provide medical IQ, and there won't be a shortage," he said on an episode of the "People by WTF" podcast published in April.
But doctors have told BI that AI can't and shouldn't replace clinicians just yet.
AI can't replicate physicians' presence, empathy, and nuanced judgment in uncertain or complex conditions, said Dr. Shravan Verma, the CEO of a Singapore-based health tech startup.
Chatbots and AI tools can handle the first mile of care, but they must escalate to qualified professionals when needed, he told BI last month.
Do you have a story to share about AI in healthcare? Contact this reporter at cmlee@insider.com.
Read the original article on Business Insider
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
21 minutes ago
- Yahoo
Project to boost cancer screening uptake for South Asian Muslim women
A University of Bradford research project has sought to increase cancer screening uptake among South Asian Muslim women in Bradford, by engaging men in their families. The three-year Muslim Families Screening For Life project, funded by Yorkshire Cancer Research, asked South Asian Muslim men to speak to women in their family about the importance of attending cancer screening appointments. The project was delivered in partnership with organisations including Faith in Communities, Bradford Institute for Health Research, Sharing Voices Bradford, and the British Islamic Medical Association. A total of 1,072 males aged 16 and over took part, and 58 workshops were held at Bradford district venues, including mosques, gyms, and community centres. Findings of the project were presented at an event held at the University of Bradford on Tuesday, June 24. Dr Mel Cooper, reader in maternity and migrant health at the University of Bradford and joint project lead, said: "This project has seen grassroots-level organisations working together with the same purpose. "We are delighted to reach 1,072 men, which was above our target of 1,000. "We hope the word has spread through the communities and more women have taken up screenings after this project and we have helped to save lives." Participants were encouraged to speak with six women in their families about the importance of attending cancer screenings. Cancer screening rates for breast, cervical, and bowel cancer among Muslim women in Bradford are thought to be relatively low: in the BD2, 3, 5, 7, and 8 postcode areas alone, 18,500 screenings are currently overdue, according to a spokesperson from the University of Bradford. Zahid Khan, a community development worker with Sharing Voices Bradford, who helped to deliver the workshops, said: "The men have been receptive. "Men have come up to us after the sessions and said the project is important and they were glad that we're talking about it." Sixth form students at Bradford schools and colleges, and students at the University of Bradford, also contributed to the research, which included questionnaires and interviews to assess knowledge and understanding of the need for the project. Dr Stuart Griffiths, director of research at Yorkshire Cancer Research, said: "The government must fund and quickly introduce innovations in screening to help reduce health inequalities so more people in Yorkshire live longer, healthier lives, free of cancer." Early findings have been analysed by the University of Bradford and will inform future projects.
Yahoo
29 minutes ago
- Yahoo
Can Fundstrat's Tom Lee Make BitMine Immersion Stock the ‘MicroStrategy of Ethereum'?
BitMine Immersion (BMNR) shares soared more than 500% on Monday after the crypto mining firm named Fundstrat's head of research Tom Lee its new chairman of the board. More importantly, the Los Angeles-headquartered company announced plans to raise about $250 million through a private placement, proceeds from which, it confirmed, will be used to buy Ethereum (ETHUSD). Jeff Bezos Unloads $5.4B in Amazon Shares: Should You Buy or Sell AMZN Stock Now? Options Flow Alert: Bulls Making Their Move in GOOGL Stock Is Palantir Stock a Buy, Sell, or Hold on New Nuclear Deal? Markets move fast. Keep up by reading our FREE midday Barchart Brief newsletter for exclusive charts, analysis, and headlines. Today's rally has inflated BMNR's market cap to well over $150 million. Under Tom Lee as chairman, BitMine plans on using Ether as its primary treasury reserve asset. According to the market veteran, ETH stands to benefit rather significantly now that the U.S. Senate has approved the much-anticipated 'Genius Act.' The legislation clears the path for stablecoins to traverse further into mainstream adoption, which could benefit the world's second-largest crypto since 'Ethereum is the blockchain where the majority of stablecoin payments are transacted,' Lee said in a statement today. Moving forward, the mining company will use ETH held per share as a key performance metric, he added. In short, Lee aims at transforming BitMine Immersion shares into the 'MicroStrategy of Ethereum.' Tom Lee has picked Ethereum as the future reserve asset for BMNR since he remains bullish on what the future holds for the cryptocurrency. According to his forecast from late 2024, ETH should hit $5,000 by the end of this year, indicating potential for a 100% rally over the next six months. However, in the near term, it's reasonable to question whether BitMine Immersion stock will be able to sustain its gains accumulated this morning. Investors should learn from the example of SharpLink (SBET) as well, which announced a similar Ether strategy last month, resulting in an explosive initial rally that all but faded just as quickly in the subsequent sessions. Caution is warranted in chasing the cosmic run in BMNR shares today also because no Wall Street analysts tracked by Barchart currently cover this BTC mining firm, indicating it's not on the radar for institutional investors – at least for now. On the date of publication, Wajeeh Khan did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
35 minutes ago
- Yahoo
New AI tool is better than doctors at diagnosing complicated medical issues, Microsoft says
Microsoft said it is one step closer to 'medical superintelligence' after a new artificial intelligence (AI) tool beat doctors at diagnosing complex medical problems. Tech giants are racing to develop superintelligence, which refers to an AI system that exceeds human intellectual abilities in every way – and they're promising to use it to upend healthcare systems around the world. For the latest experiment, Microsoft tested an AI diagnostic system against 21 experienced physicians, using real-world case studies from 304 patients that were published in the New England Journal of Medicine, a leading medical journal. The AI tool correctly diagnosed up to 85.5 per cent of cases – roughly four times more than the group of doctors from the United Kingdom and the United States, who had between five and 20 years of experience. Related As AI reshapes patient care, human nurses are pushing back against its creeping influence The model was also cheaper than human doctors, ordering fewer scans and tests to reach the correct diagnosis, the analysis found. Microsoft said the findings indicate that AI models can reason through complex diagnostic problems that stump physicians, who specialise in their fields but are not experts in every aspect of medicine. However, AI 'can blend both breadth and depth of expertise, demonstrating clinical reasoning capabilities that, across many aspects of clinical reasoning, exceed those of any individual physician,' Microsoft executives said in a press release. 'This kind of reasoning has the potential to reshape healthcare'. Microsoft does not see AI replacing doctors anytime soon, saying the tools will instead help physicians automate some routine tasks, personalise patients' treatment, and speed up diagnoses. Microsoft's AI system made diagnoses by mimicking a doctor's process of collecting a patient's details, ordering tests, and eventually narrowing down a medical diagnosis. A 'gatekeeper agent' had information from the patient case studies. It interacted with a 'diagnostic orchestrator' that asked questions and ordered tests, receiving results from the real-world workups. Related Is AI going to steal your job? Probably not, new study finds The company tested the system with leading AI models, including GPT, Llama, Claude, Gemini, Grok, and DeepSeek. OpenAI's o3 model, which is integrated into ChatGPT, correctly solved 85.5 per cent of the patient cases, compared to an average of 20 per cent among the group of 21 experienced doctors. The researchers published their findings online as a preprint article, meaning it has not yet been peer-reviewed. Microsoft also acknowledged some key limitations, notably that the AI tool has only been tested for complicated health problems, not more common, everyday issues. Related AI to shape EU health policymaking without new rules The panel of doctors also worked without access to their colleagues, textbooks, or other tools that they might typically use when making diagnoses. 'This was done to enable a fair comparison to raw human performance,' Microsoft said. The company called for more real-world evidence on AI's potential in health clinics, and said it will 'rigorously test and validate these approaches' before making them more widely available.
