science & techanalysisThe Conversation
Dr. ChatGPT is getting remarkably good at diagnosing health problems - but actual doctors are still better at weighing treatment options
AI chatbots can now match or beat doctors at diagnosis: one 2026 study found OpenAI's o1 model reached 78% accuracy on complex cases from the New England Journal of Medicine. Doctors build illness scripts, mental shortcuts of what a disease typically looks like, and large language models work in a similar way by predicting patterns learned from medical text. Management is harder. After a diagnosis, doctors choose among several reasonable options, such as surgery, radiation or monitoring a slow-growing prostate cancer, and the right choice depends on the patient's other conditions, values and tolerance for uncertainty. Scoring tools can handle the numbers, but risk and trust are shaped by individual experience, which AI cannot fully weigh.
People now ask chatbots about symptoms, so it matters where those tools are reliable. Diagnosis is increasingly automated, while judgment about treatment remains a human task.
Illustrative photo · Unknown author, CC BY 3.0, via Wikimedia Commons
