Investigating the Potential Effects of Medical AI Systems on Physician Autonomy: Pretest of a Semistructured Qualitative Interview Guide.

Background: AI is an increasingly prominent feature of contemporary health care, with medical AI systems beginning to support diagnostic and therapeutic processes in many clinical domains. Alongside the anticipated benefits of these technologies, their introduction also raises broader questions about how clinical work and professional roles may change. In particular, medical AI systems may affect physician autonomy, a key factor influencing the acceptance and long-term implementation of new medical technologies. Objective: The aim of this study was to develop and pretest a semistructured interview guide concerning the potential effects of medical AI systems on physician autonomy. Methods: The interview guide was theoretically grounded in a 7-component model of physician autonomy proposed by Schulz and Harrison. Semistructured qualitative interviews were conducted with a sample of 7 hospital physicians. Interview recordings were transcribed and analyzed using a hybrid inductive-deductive thematic approach: themes were first identified inductively from participant responses and subsequently mapped onto the 7-component model of physician autonomy proposed by Schulz and Harrison. Data were analyzed to assess both the potential effects of medical AI systems on physician autonomy and the methodological adequacy of the interview guide. Results: Most participants did not express strong concerns about losing clinical autonomy through the introduction of AI systems. However, several autonomy-related risks were identified, including potential deskilling, automation bias, limited system explainability, and increasing economic or cost-related pressures. Participants emphasized that AI should serve as a supportive tool rather than a substitute for physician judgment. All physicians agreed that AI systems should not replace clinicians as primary clinical decision-makers. Conclusions: Medical AI was largely viewed as compatible with physician autonomy; however, participants highlighted important risks that warrant attention in future research and system design. Our preliminary findings suggest that autonomy-related concerns extend beyond the direct loss of decision-making authority and include broader professional, cognitive, and organizational dimensions. However, our inductively identified themes and subthemes did not fully reflect all components of physician autonomy, indicating the need for further refinement of how to assess physician autonomy in qualitative research.