Acceptability, Perceptions, Trust, and Ethical Issues of Generative Artificial Intelligence in the Patient–Clinician Relationship: Exploratory Mixed Methods Study of Patients and General Practitioners (Preprint)

BACKGROUND The integration of generative artificial intelligence (GenAI) into health care raises important questions about acceptability, trust, and the patient–clinician relationship. Although interest in these technologies is increasing, their adoption by patients and general practitioners remains uncertain and conditional. OBJECTIVE This study aimed to examine psychosocial factors associated with the acceptability of GenAI in health care by comparing the perceptions, expectations, and concerns of patients and general practitioners. METHODS We conducted an exploratory mixed methods study in France. A total of 13 patients and 9 general practitioners completed questionnaires based on an adapted Technology Acceptance Model (TAM) and a technophilia and technophobia scale. Quantitative data were analyzed using comparative and correlational analyses. Semistructured interviews with 10 patients and 5 physicians were analyzed using thematic analysis. RESULTS Perceived usefulness was strongly associated with intention to use (r = .76, P<.001), whereas perceived ease of use was not (r = .11, P=.629). Physicians rated GenAI as easier to use than patients did (t(20) = 2.63, P=.016). Associations involving technophilia and technophobia did not remain significant after Bonferroni correction. The interviews emphasized the importance of trust, ethical considerations, and preserving the patient–clinician relationship. Physicians were more concerned about liability and professional expertise, whereas patients emphasized human contact and data confidentiality. CONCLUSIONS Participants showed cautious and conditional receptiveness to GenAI. Its acceptability in health care involves cognitive and relational assessments of clinical usefulness, trust, accountability, and values in care. These findings highlight the need for ethical oversight, practical guidance, and human-centered implementation.