Microsoft claims its AI Diagnostic Orchestrator outperformed 21 doctors, got 85.5% of diagnoses right
Microsoft has introduced a new artificial intelligence (AI) system that it says can diagnose some of the most difficult medical cases more accurately and at a lower cost than human doctors.
The system, called the Microsoft AI Diagnostic Orchestrator (MAI-DxO), was tested using case studies published by the New England Journal of Medicine (NEJM). These cases are known for being particularly complex and usually involve teams of specialists. According to Microsoft, the AI got the correct diagnosis 85.5 per cent of the time, compared to just 20 per cent for a group of experienced doctors from the US and UK.
As more people turn to digital tools for medical advice, the company says it sees over 50 million health-related searches every day across its services like Bing and Copilot.
To test the system, Microsoft created a new challenge called the Sequential Diagnosis Benchmark (SD Bench), based on 304 real NEJM cases. The cases were turned into step-by-step scenarios, where the AI or a human doctor could ask questions or order tests before making a diagnosis. Each test had a virtual cost, helping to measure both accuracy and how wisely resources were used.
Microsoft tested several top AI models, including OpenAI's o3, Claude, Gemini, Llama, and DeepSeek, both alone and as part of MAI-DxO. The orchestrator system works by combining different models to act like a team of doctors, sharing ideas and narrowing down possible diagnoses. The best results came from using MAI-DxO with OpenAI's o3 model, the tech giant stated.
Reportedly, the results showed that the AI not only diagnosed more cases correctly but also did so with fewer and more cost-effective tests than the doctors involved in the study.
However, Microsoft admitted the research has its limits. The tests focused on rare and complex cases, not everyday health problems. Also, the doctors were not allowed to use any support tools like books or the internet during the test, unlike in real-world situations where such resources are often used.
Other tools developed by the company include RAD-DINO, which helps improve radiology processes, and Dragon Copilot, a voice assistant for doctors.
Microsoft says it is now working with health organisations to test its AI in real clinics and hospitals. Before any wider use, the technology will need to meet safety standards and get approval from regulators.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
India Today
24 minutes ago
- India Today
Mark Zuckerberg announces Meta Superintelligence Labs and 11 new hires from OpenAI, Anthropic and Google
Meta CEO Mark Zuckerberg on Monday announced the creation of Meta Superintelligence Labs (MSL), a new division aimed at developing artificial general intelligence (AGI) – an AI system that can match or exceed human intelligence across a wide range of tasks. The new group will reportedly bring together Meta's foundational AI model teams, product teams, and the company's well-known Fundamental AI Research (FAIR) division. A new lab under MSL will also focus on building the next generation of large language models (LLMs).advertisementAt the helm of this ambitious project is Alexandr Wang, the former CEO of data-labelling startup Scale AI. Wang will serve as Meta's chief AI officer, a newly created role. In a memo to employees obtained by Bloomberg, Zuckerberg described Wang as 'the most impressive founder of his generation.' Wang will be joined by Nat Friedman, who was the former GitHub CEO; he will co-lead the division. Friedman's focus will be on developing AI products and applied research. Zuckerberg said the two leaders will steer Meta's efforts to reach the next frontier in artificial intelligence.'As the pace of AI progress accelerates, developing superintelligence is coming into sight,' Zuckerberg reportedly wrote in his internal note to staff. 'I believe this will be the beginning of a new era for humanity, and I am fully committed to doing what it takes for Meta to lead the way.'advertisement According to reports, Zuckerberg has been personally leading the hiring campaign for MSL. He apparently offered generous packages and hosted many of the potential hires at his home. However, OpenAI CEO Sam Altman's claims of Meta offering a $100 million signing bonus were dismissed by the Meta company recently invested $14.3 billion in Scale AI, bringing in some of its staff, including Wang. Meta has also been in talks with AI startups like Perplexity AI and Runway, and is expected to acquire PlayAI, a small company focused on voice replication using the most talked-about aspect of this overhaul was Meta poaching people from its competitors like OpenAI, Anthropic, and Google. According to a report by Bloomberg, Meta has hired 11 top-tier AI researchers from its rivals. Wired has listed all the new hired that Zuckerberg's company has done for MSL, which includes Trapit Bansal, who was the co-creator of OpenAI's o-series models, Shuchao Bi, who was involved in GPT-4o voice and multimodal post-training, Huiwen Chang, who is known for developing image generation tools at Google Research, Ji Lin, who was a contributor to multiple GPT-4 models and reasoning systems, Joel Pobar, who was with Meta earlier and has just returned from Anthropic after a short stint, along with Jack Rae, Johan Schalkwyk and Pei Sun from Google, and Hongyu Ren, Jiahui Yu and Shengjia Zhao from launch of the superintelligence labs reflects Zuckber's long-term commitment to AI. And Meta isn't alone. The move echoes industry-wide trends. This year alone, Microsoft spent $650 million acquiring much of Inflection AI's team, while Amazon has hired key staff from Adept. - Ends
The Hindu
41 minutes ago
- The Hindu
Apple weighs using Anthropic or OpenAI to power Siri in major reversal: Report
Apple is weighing using artificial intelligence technology from Anthropic or OpenAI to power a new version of Siri, instead of its own in-house models, Bloomberg News reported on Monday. The iPhone maker has had discussions with both companies about using their large language models for Siri, asking them to train versions of their LLMs that could run on Apple's cloud infrastructure for testing, the report said, citing people familiar with the discussions. Apple's investigation into third-party models is at an early stage and the company has not made a final decision on using them, the report said. Amazon-backed Anthropic declined to comment, while Apple and OpenAI did not respond to Reuters requests. The company had said in March that AI improvements to its voice assistant Siri will be delayed until 2026, without giving a reason for the setback. Apple shook up its executive ranks to get its AI efforts back on track after months of delays, resulting in Mike Rockwell taking charge of Siri, as CEO Tim Cook lost confidence in AI head John Giannandrea's ability to execute on product development, Bloomberg had reported in March. At its annual Worldwide Developers Conference earlier this month, Apple focused more on incremental developments that improve everyday life — including live translations for phone calls — rather than the sweeping ambitions for AI that Apple's rivals are capitalizing. Apple software chief Craig Federighi had then said it is opening up the foundational AI model that the iPhone maker uses for some of its own features to third-party developers, and that the company will offer both its own and OpenAI's code completion tools in its key Apple developer software.
Economic Times
an hour ago
- Economic Times
ChatGPT to be closed as OpenAI bleeds talent to Meta? Here's what OpenAI Chief Research Officer Mark Chen reveals
OpenAI will shut its doors for a week to help burned-out employees rest, Wired reports. After months of 80-hour weeks, staff are being told to step away from their screens. Only top executives will stay on duty. The break is unusual. OpenAI rarely halts its intense push for artificial general intelligence (AGI). 'Meta knows we're taking this week to recharge and will take advantage of it to try and pressure you to make decisions fast and in isolation,' Chief Research Officer Mark Chen warned in a Slack memo, as seen by Wired. This break lands at a tense moment. Meta's CEO Mark Zuckerberg has already hired away seven researchers from OpenAI. Among them are Lucas Beyer, Alexander Kolesnikov, Xiaohua Zhai and Trapit Bansal, a key contributor to OpenAI's o1 wants OpenAI's best. And it knows exactly when to strike. 'Meta knows we're taking this week to recharge and will take advantage of it to try and pressure you to make decisions fast and in isolation,' Chen repeated in the same offers are no small talk. Sam Altman, OpenAI's CEO, told his brother Jack on a podcast that Meta has dangled signing bonuses over $100 million for some staff. 'Over the past month, Meta has been aggressively building out their new AI effort, and has repeatedly (and mostly unsuccessfully) tried to recruit some of our strongest talent with comp-focused packages,' Chen wrote on Slack. A source close to Meta told Wired the company wants the best from OpenAI and Google. 'They haven't necessarily expanded the band, but for top talent, the sky is the limit,' the source Beyer, who recently left OpenAI for Meta Platforms, has denied claims that he was lured with a huge payout. In a direct post on X (formerly Twitter), he wrote, 'Hey all, couple quick notes: 1) Yes, we will be joining Meta. 2) No, we did not get 100M sign-on, that's fake news. Excited about what's ahead though, will share more in due time.'Beyer's response came after OpenAI CEO Sam Altman said on the Uncapped podcast that Meta made 'giant offers' to poach his people. 'Like $100 million signing bonuses. It is crazy,' Altman said on air. He also claimed he was relieved people had turned the offers down. Beyer is not the only OpenAI researcher jumping ship. Alexander Kolesnikov and Xiaohua Zhai have joined him, leaving OpenAI's Zurich office for Meta's expanding AI push. According to The Information , the list is growing. Shengjia Zhao, Jiahui Yu, Shuchao Bi and Hongyu Ren are also said to be joining Mark Zuckerberg's superintelligence effort impact runs deep. Cheng Lu, a technical staff member at OpenAI, posted on X, 'Not too many people outside the company know how talented and hardcore they are.' He added, 'Such a huge loss to OpenAI and I feel really disappointed that the leadership didn't keep them.' His post went viral before he deleted it. — kimmonismus (@kimmonismus) OpenAI staff admit the fatigue is real. Inside, some feel Meta is pouncing while they're exhausted. 'If they pressure you, or make ridiculous exploding offers just tell them to back off, it's not nice to pressure people in potentially the most important decision,' one research leader wrote to staff, as Wired says OpenAI is adjusting pay and reward structures. 'I feel a visceral feeling right now, as if someone has broken into our home and stolen something,' Chen wrote. He added, 'Please trust that we haven't been sitting idly by.' He promised he and Altman are 'around the clock to talk to those with offers.'But Chen made clear he has limits. 'While I'll fight to keep every one of you, I won't do so at the price of fairness to others,' he may rest for a week, but leaders will stay working. Some inside believe the company has focused too much on launching new features. Chen wants them to return to the core goal. 'We need to remain focused on the real prize of finding ways to compute (a lot more supercomputers are coming online later this year) into intelligence,' he wrote. 'This is the main quest, and it's important to remember that skirmishes with Meta are the side quest.'Sam Altman backed him up in Slack. 'It's been really amazing to watch Mark's leadership and integrity through this process, especially when he has had to make tough decisions. Very grateful we have him as our leader!'OpenAI has weathered crisis before, including Altman's brief firing last year. But the stakes now feel different. Some see the break as vital care. Others see it as damage exhausted researcher posted, then deleted, a simple truth: this mission is not just code. It's people. And people burn OpenAI asks staff to rest, Meta waits. And the world's AI race grinds on — powered by human brains that need a moment to breathe.
