{"id":1911534,"date":"2026-04-30T18:40:05","date_gmt":"2026-04-30T15:40:05","guid":{"rendered":"https:\/\/analyse.optim.biz\/?p=1911534"},"modified":"2026-04-30T18:40:05","modified_gmt":"2026-04-30T15:40:05","slug":"study-ai-can-outperform-doctors-on-diagnosing-cases","status":"publish","type":"post","link":"https:\/\/analyse.optim.biz\/?p=1911534","title":{"rendered":"Study: AI can outperform doctors on diagnosing cases"},"content":{"rendered":"

[analyse_image type=”featured” src=”https:\/\/helios-i.mashable.com\/imagery\/articles\/05wVA0W3rOMt8cQjgYKDLyu\/hero-image.fill.size_1200x675.v1777502088.jpg”]<\/p>\n

\n
\n
\n
\n
\n
\n
Home<\/a>><\/span>Tech<\/a><\/div>\n<\/section>\n<\/div>\n<\/div>\n

AI can reason like a doctor, study says<\/h1>\n
Researchers tested AI on medical scenarios and were surprised by what they found.<\/div>\n
\n
\n \u00a0By\u00a0<\/p>\n
Rebecca Ruiz<\/a><\/p>\n
\n
\n
\n
\"Rebecca<\/p>\n
\n
Rebecca Ruiz<\/div>\n
Senior Reporter<\/div>\n<\/div>\n<\/div>\n<\/div>\n
\n Rebecca Ruiz is a Senior Reporter at Mashable. She frequently covers mental health, digital culture, and technology. Her areas of expertise include suicide prevention, screen use and mental health, parenting, youth well-being, and meditation and mindfulness. Rebecca’s experience prior to Mashable includes working as a staff writer, reporter, and editor at NBC News Digital and as a staff writer at Forbes. Rebecca has a B.A. from Sarah Lawrence College and a masters degree from U.C. Berkeley’s Graduate School of Journalism.\n <\/div>\n

Read Full Bio<\/a><\/div>\n<\/div>\n<\/div>\n

\u00a0on\u00a0

Share on Facebook<\/span><\/a>Share on Twitter<\/span><\/a>Share on Flipboard<\/span><\/a><\/div>\n<\/div>\n<\/header>\n<\/div>\n
\n
\"Doctor<\/p>\n
AI performed as well or better than physicians in new study. <\/span>Credit: Rawlstock via Moment \/ Getty Images<\/span><\/div>\n<\/div>\n
\n
<\/div>\n<\/div>\n
\n

Artificial intelligence<\/a> that can “reason” is now capable of diagnosing real-life medical scenarios as well as or better than physicians, according to the results of a study<\/a> published Thursday in Science.<\/em><\/p>\n

The researchers used previously unknown clinical cases to test OpenAI<\/a>‘s reasoning model o1 against the company’s older model, GPT-4, as well as physicians and medical residents in training. <\/p>\n

In a range of experiments, the o1 model often improved significantly on GPT-4’s diagnostic ability and bested physicians, too. When tested with the electronic health records of random emergency department cases from a Boston hospital, the o1 model was diagnostically accurate more than two-thirds of the time at initial triage. Two expert attending physicians had correct diagnoses roughly half of the time. <\/p>\n

\n
\n
You May Also Like<\/strong><\/p>\n
<\/div>\n
<\/div>\n
SEE ALSO:<\/span>What AI can tell you about your blood test<\/span><\/a><\/div>\n

Dr. Robert Wachter, professor and chair of the Department of Medicine at the University of California, San Francisco, described the study’s findings as “important” and suggested it’s now “indisputable” that modern AI will outperform older large language models and doctors when asked to identify the right diagnosis and next step. He was not involved in the study. <\/p>\n

However, Wachter, author of “A Giant Leap: How AI is Transforming Healthcare and What That Means for Our Future<\/u><\/em><\/a>,” added that more research is necessary before AI is fully implemented in clinical practice. <\/p>\n

“The question is how closely this replicates real life, and the answer is moderately well but not perfectly,” Wachter wrote in an email. <\/p>\n

As the study’s authors acknowledge, the experiments were limited to text-only input and didn’t include the visual and auditory clues and cues that doctors often rely on for diagnosis. These can include a patient’s level of distress and medical imaging. <\/p>\n

“GenAI can probably begin to integrate these inputs but for now, a test of a written, and often artificially ‘clean’ clinical case scenario is not the same as going into an ER and dealing with the chaos,” Wachter said. “Just watch The Pitt<\/em><\/a>.”<\/p>\n

\n
Mashable Light Speed<\/span><\/div>\n
\n
\n
Want more out-of-this world tech, space and science stories?<\/div>\n
Sign up for Mashable’s weekly Light Speed newsletter.<\/div>\n
\n By clicking Sign Me Up, you confirm you are 16+ and agree to our Terms of Use<\/a> and Privacy Policy<\/a>.\n <\/div>\n<\/div>\n
\n
Thanks for signing up!<\/div>\n<\/div>\n<\/div>\n<\/section>\n
SEE ALSO:<\/span>When is ‘The Pitt’ Season 3 coming out?<\/span><\/a><\/div>\n

Can AI replace a doctor? <\/h2>\n

Dr. Ashwin Ramaswamy, an instructor of urology at Mt. Sinai who has studied AI’s ability to respond to consumer health inquiries, shared a similar response to the study. <\/p>\n

While he commended the study’s design, Ramaswamy noted that the AI reasoned over clinical information that had been collected, filtered, documented by humans. In real life, patients may be afraid, intoxicated, or actively deteriorating, among other challenges physicians encounter when making diagnoses.<\/p>\n

“This is valuable and it shows the progress of the technology that it performed so well, but it skips a central part of the job of ‘being a doctor,'” Ramaswamy said in an email. <\/p>\n

He also wished for specific details about the errors made both by physicians and the LLM. If the model made an understandable near-miss, that’s different than a dangerous, unexplainable mistake. <\/p>\n

In Ramaswamy’s own recent evaluation of ChatGPT Health, published as a peer-reviewed advance paper in Nature Medicine<\/em><\/a>, he and fellow researchers found that AI’s failure modes can be “jagged.” In other words, AI might perform well when diagnosing a rare, difficult disease, but still miss something clinically obvious. <\/p>\n

\n
\n
\n
Related Stories<\/div>\n
\n<\/div>\n