Models & Research
Two studies find AI chatbots still unreliable for medical diagnosis and triage
Oxford researchers found chatbot users identified their condition correctly in about a third of cases, while a Mount Sinai study said ChatGPT Health under-triaged over half of emergencies.