ChatGPT has the ability to diagnose patients in emergency centers in the future, a new pilot study that analyzed how this large-scale artificial intelligence (AI)-based language model could be used to support doctors showed.
The research, which was published in the journal Annals of Emergency Medicine, found that the chatbot diagnosed patients at a similar level to trained doctors. These findings will be presented this weekend at the European Congress of Emergency Medicine.
Researchers from the Netherlands’ Jeroen Bosch Hospital entered doctors’ notes, such as examination results, symptoms and laboratory results, along with anonymous information about about 30 patients, into two versions of ChatGPT. They found that the list of possible diagnoses of the doctors in the emergency centers and those of the chatbot overlapped in about 60 percent. “We found that ChatGPT had good success in generating a list of possible diagnoses and suggesting the most likely option. We also found a large overlap with doctors’ lists of possible diagnoses. “Simply put, this shows that ChatGPT was able to suggest medical diagnoses much like a doctor would,” study author Dr. Heide ten Berg, from the Department of Emergency Medicine at Jeroen Bosch Hospital, said in a statement.
Source: EuroNews