Study: Chat GPT is more empathetic than doctors

Can a chatbot assistant answer patient questions via artificial intelligence with the same quality and empathy as if the answers came from actual doctors? With this question in mind, a recent study has been published in the journal Jama International Medicine. The answer? Yes, the chatbot is even better.

The study randomly selected 195 patient questions posted on the social media platform Reddit’s forum, then had them answered in writing by doctors and the much-publicized AI chatbot Chat GPT. When licensed healthcare professionals examined the responses in terms of the quality of the information provided and empathy in the response, they found that the chatbot’s responses were preferred over the doctor’s in 78.6% of the 585 evaluations.

The chatbot received a 3.6 times higher prevalence of good or very good ratings in the quality of its answers. Furthermore, it proved superior in behaviour – with a 9.8 times higher score in empathy.

Chat GPT also had more to say. The average length of the doctor’s answers was 52 words, while the chatbot reached an average of 211 words.

The study authors suggest that the technology should be further explored, which could, for example, lead to the chatbot writing answers that doctors are able to edit. “Randomized tests could further assess whether using AI assistants can improve responses, reduce physician burnout, and improve patient outcomes,” the report says.