Understanding the Use of AI and Ethics Through the Lens of Embodiment : A Practice-Oriented Critique to the Existing AI Research in Healthcare Beyond Theoretical Perspectives

Sammanfattning

Motivation: Healthcare is currently undergoing a great transformation as Artificial Intelligence (AI) becomes increasingly integrated into diagnostic processes. Sweden, and its given AI readiness, offers a particularly suitable context for investigation. While AI is promising diagnostic accuracy and efficiency, it also challenges traditional forms of medical reasoning, raises ethical concerns, and introduces tensions between human expertise and algorithmic logic. Understanding how physicians experience and respond to AI in diagnostic contexts is crucial, both for responsible technology development and also for maintaining patient-centered, ethically grounded care in an evolving healthcare landscape. Purpose: This study aims to explore how AI influences clinical decision-making in healthcare diagnostics. A particular focus is directed towards embodied expertise and the ethics of care. It investigates how physicians rely on their intuition, sensory judgment, and relational understanding when navigating diagnostic tasks. The study also considers how these insights contribute to a broader understanding of the ongoing integration of AI into the healthcare system. Methodology: The research follows a qualitative, exploratory design using semi-structured interviews with nine practicing physicians in Sweden. Participants were selected through purposive sampling to ensure a diversity of clinical backgrounds and experiences. A thematic coding table was used to identify recurring patterns and tensions in how physicians perceive, use, and relate to AI in diagnostic decision-making. The study is grounded in the theoretical frameworks of embodiment perspective and ethics of care. Results: The findings highlight the irreplaceable role of embodied expertise in healthcare diagnostics and emphasize physicians’ preference for human-AI collaboration over automation. While AI is valued for administrative support and specific clinical tasks, concerns remain around ethics, trust, and contextual sensitivity, particularly in complex decision-making.