One of the discussions about ChatGPT its role in medicine. Now experiment with medical professionals discovered that this artificial intelligence beats doctors when providing answers empathic and high quality written questions from patients.

While artificial intelligence (AI) will not replace a doctora JAMA Internal Medicine article suggests that the collaboration of professionals with technologies such as ChatGPT could “revolutionize medicine,” says the University of California, San Diego, responsible for the study.

Directed by John W. Ayers, studio compared V written responses what doctors and ChatGPT gave to patients’ questions. According to the results, group professionals who evaluated them preferred answers AI He 79% of the timeand categorized them as better quality and more empathic.

Interested in learning how to make predictive models more clinically relevant? Find out how to do it at @CritCareMed!

Via Twitter @WardiGabriel

The team set out to find out if ChatGPT could accurately answer the questions patients send to their doctors.

If so, the researchers say, AI models could integrate V systems sanitary improve doctors’ responses to patients’ questions, especially since the covid-19 pandemic has accelerated virtual care, thus easing the growing workload on doctors.

To obtain a large and diverse sample of non-personally identifiable questions and answers from physicians, the team turned to “Askdox” from the platform reddit forums.

In it, users publish questions that are answered by verified medical professionals. While anyone can answer a question, moderators check the credibility of professionals, and responses indicate the level of credibility of the respondent, the university said in a statement.

The team randomly selected 195 AskDocs exchanges in which a verified doctor answered a public question. He provided the original ChatGPT question and asked for an answer.

Group three professionals health professionals evaluated each question and the corresponding answers without knowing whether it came from a doctor or from ChatGPT. they compared answers depending on the quality belonging information And sympathyand indicate which one they prefer.

Group of appraisers preferred V answers from ChatGPT for doctors in 79% of the time.

“The ChatGPT messages contained detailed and accurate information, often covering more aspects of patients’ questions than doctors’ answers,” said study co-author Jessica Kelly.

Except quality answers ChatGPT was significantly superiors to one of doctors: the responses were of good or very good quality 3.6 times higher for ChatGPT

answers they were also more empathetic: empathic or very empathic responses were 9.8 times higher for AI.