Examining patient and provider perspectives on AI communication tools in healthcare

Laura Chang – School of Medicine – Stanford University; David Magnus – Stanford Center for Biomedical Ethics – Stanford University

Abstract: As artificial intelligence tools proliferate in healthcare, researchers have turned to generative AI to support clinicians in tasks that can contribute to burnout. One set of tasks involves after-hours EHR work, which includes responding to patients’ electronic messages. In response, some healthcare systems have deployed AI-mediated communication tools within EHRs to help clinicians draft messages and respond to patients. Evaluations of these tools focus on providers’ perspectives and the content of the messages provided by generative AI.

We argue, however, that such evaluations are incomplete as they neglect to include patients’ perspectives on the implementation of AI communication tools into the patient-provider relationship. Therefore, we present findings from an interview study aimed at understanding patients’ perceptions of AI-mediated communication tools in healthcare and comparing such perceptions to those of providers’. Preliminary findings include disparate preferences surrounding the disclosure of AI messaging use, perceptions of trust in the patient-provider relationship given AI messaging use, and concerns regarding the loss of human touch in medicine. We conclude with a discussion about the consequences of excluding patient perspectives from AI implementation practices in healthcare and present potential hypotheses regarding the impact of AI tools on patient-provider communication.

Keywords: patient-provider communication, artificial intelligence in healthcare, AI-mediated communication tools