Add to Favourites
To login click here
In a pilot study conducted by researchers from the department of emergency medicine and clinical chemistry and haematology at Jeroen Bosch Hospital in the Netherlands, the artificial intelligence chatbot ChatGPT performed as well as a trained doctor in suggesting likely diagnoses for patients being assessed in emergency medicine departments. The study found that the shortlists generated by ChatGPT had a large overlap (around 60%) with the shortlists made by emergency medicine doctors, and ChatGPT had the correct diagnosis within its top five likely diagnoses in 97% of the cases.
