Ethics in patient preferences for artificial intelligence–drafted responses to electronic messages

Ethics in patient preferences for artificial intelligence–drafted responses to electronic messages
JAMA Network Open; Joanna S. Cavalier, Benjamin A. Goldstein, Vardit Ravitsky, Jean-Christophe Bélisle-Pipon, Armando Bedoya, Jennifer Maddocks, Sam Klotman, Matthew Roman, Jessica Sperling, Chun Xu, Eric G. Poon, Anand Chowdhury; 3/25
The rise of electronic communication sent to clinicians via the patient portal has directly led to clinician burnout and dissatisfaction. With patients increasingly messaging their clinicians, replying to in-basket messages (akin to email) has become a burdensome task consisting of medical questions, refill requests, and administrative and scheduling requests. This survey study of 1,455 respondents showed that while overall satisfaction was high (>75%) regardless of author, respondents preferred responses written by AI [artificial intelligence] over those written by a human ... However, when an AI author was disclosed, satisfaction was lower for AI compared with a human author ... Meaning: Reduced satisfaction due to AI disclosure should be balanced with the importance of patient autonomy and empowerment.