AI and language in healthcare: The role of doctor identity and linguistic alignment in perceptions of doctor quality

AI will play an increasingly large role in providing healthcare services in the near future, creating a pressing need to examine how people perceive AI compared to humans in healthcare. To this end, this study examines how people perceive AI and human doctors in a common medical scenario of giving healthcare advice over an online chat platform. Participants read an online chat conversation that featured a patient asking for medical advice with either an AI or human doctor. In addition, the doctor either linguistically aligned with the patient (by repeating medical terms previously used by the patient) or not (by using synonyms of the patient's terms). Linguistic alignment in conversation has been found to increase positive perceptions of interactions and therefore may be a useful tool to increase patient satisfaction in healthcare interactions. AI doctors but not human ones were penalized when not linguistically aligning with the patient, even though these doctors gave the same exact medical advice. Additionally, across a range of measures, participants in general had more negative perceptions of AI doctors compared to human ones. Overall, these results suggest that people hold AI to different standards as well as judge AI harsher than humans in medical settings.