Study: ChatGPT beat doctors for quality and empathy

As reported here. 

"In JAMA Internal Medicine, a new report compared the quality and empathy of responses to patient questions for doctors vs ChatGPT, a generative AI model that has already been superseded by GPT4. [...] The answers were reviewed by a panel of 3 health care professionals blinded to whether the response was from a doctor or the chatbot. [...] For quality, the evaluators preferred the ChatGPT response 79% of the time; the quality score of good and very good was 79% for the chatbot vs 22% for physicians, a near 4-fold difference. That gap was even greater for empathy rating of the responses, nearly 10-fold higher favoring the chatbot, for proportion of empathetic or very empathic responses (45.1 vs 4.6%)."