Physicians’ Experiences of Trust, Ethics, and Identity in AI-Based Clinical Decision Support Systems

Abstract

Artificial intelligence (AI) has emerged as a transformative force in healthcare, reshaping diagnostic processes, treatment planning, and patient management. Within this broad landscape, little is known about how physicians experience the psychological, ethical, and professional dimensions of engaging with AI-based clinical decision support systems (AI-CDSS). Despite rapid adoption, existing research has focused primarily on technical performance, leaving unanswered questions about how physicians negotiate trust, autonomy, and accountability in everyday clinical decision-making. This study specifically asks: how do physicians construct, interpret, and manage their sense of trust, ethical responsibility, and professional identity when interacting with AI-CDSS in clinical practice? Here, we apply an interpretative phenomenological approach to examine physicians’ lived experiences with AI-CDSS and to explore the meanings they ascribe to this evolving collaboration. Semi-structured interviews were conducted with 18 physicians across multiple specialties, and transcripts were analyzed thematically using interpretative phenomenological analysis. The findings reveal four key themes: ambivalent trust in AI recommendations, emotional strain and identity challenges, ethical dilemmas in shared decision-making, and adaptive strategies for integrating AI into practice. Physicians reported valuing the efficiency of AI while struggling with feelings of diminished autonomy and heightened moral responsibility. Direct quotations illustrate how the presence of AI in clinical practice is experienced not as neutral support but as a transformative influence on the essence of professional responsibility. Overall, the study sharpens the focus on the lived question of “what it means to be a physician in the age of AI,” advancing current understanding by highlighting that the success of AI integration depends not only on technological precision but also on human-centered factors of trust, identity, and ethics. The study underscores the importance of phenomenological inquiry for future research and suggests that sustainable AI adoption must integrate both technical and experiential dimensions of medical practice.