A small US study is testing whether AI chatbots can safely suggest possible diagnoses for patients before they are seen by a clinician, potentially helping triage cases ahead of GP appointments. According to reporting on the research, the AI’s suggested diagnoses match the eventual confirmed diagnosis in more than half of cases.
Outlets frame the work as an early but potentially important step toward using AI in routine care. The emphasis differs: one source highlights expert views describing the approach as “crucial,” while another focuses on doctors actively testing the concept in a controlled setting. Both accounts describe the project as an investigation rather than a ready-to-deploy diagnostic tool.
The study’s findings are presented in the context of assessing safety and usefulness, with the goal of understanding how such systems might support—rather than replace—clinical decision-making during patient intake and referral to GPs.
