Algorithmic gaze and subject occlusion: a medical ethical critique of artificial intelligence diagnosis and treatment from a foucaultian perspective.

This paper develops a Foucauldian analysis of artificial intelligence (AI) in medical diagnosis, arguing that AI introduces an "algorithmic gaze" that reshapes the epistemic and moral structure of clinical practice. By converting illness narratives into data while generating forms of simulated empathy, AI systems reorder how patients become visible, intelligible, and governable within medical discourse. These developments produce three structural tensions: first, algorithmic opacity alters the conditions under which informed consent, contestation, and responsibility can meaningfully operate; second, simulated empathy appears to acknowledge suffering while subtly directing patients' expressive possibilities; and third, the delegation of emotional engagement to AI fragments the unity of care traditionally embodied in the clinician-patient relationship.Rather than treating these shifts as technologically inevitable, the paper argues that their ethical significance depends on institutional design. Four directions for reconstruction are proposed: securing identifiable responsibility and human-overridable decision pathways; integrating algorithmic outputs into dialogical clinical communication; adopting participatory data governance and strengthening algorithmic literacy; and grounding AI development in a renewed conception of medical humanism that guides design from the outset. The analysis concludes that AI can support, rather than erode, the moral foundations of medicine only if embedded within institutional arrangements that sustain autonomy, dignity, and relational understanding.