From Zadig to Artificial Intelligence: Clinical Observation, Epistemic Fragility, and the Future of Medical Reasoning

ABSTRACT: Background: The rapid integration of artificial intelligence (AI) into clinical practice has revived longstanding debates about the foundations of diagnostic reasoning. While AI systems rely on large datasets derived from past patients, the clinician’s craft remains rooted in direct observation, sensory examination, and contextual interpretation. Objective: To trace the intellectual lineage connecting Voltaire’s Zadig, the evolution of the clinical method, the deductive tradition exemplified by Sherlock Holmes, and the contemporary challenges posed by AI in medicine. Discussion: Voltaire’s Zadig anticipates the logic of modern clinical reasoning through its emphasis on inference from subtle signs. This evidential paradigm reappears in the anatomic and clinical revolution, bacteriology, and evidence‑based medicine, and is later embodied in Arthur Conan Doyle’s Sherlock Holmes, whose method was explicitly modelled on clinical diagnosis. AI introduces new fundamental risks, including algorithmic affixing, loss of sensory skills, and the perpetuation of historical preconceptions. The internist’s observational abilities remain essential for generating primary data and contextualizing algorithmic outputs. Conclusion: The future of medicine requires harmonizing traditional clinical skills with AI‑driven tools. The internist must remain a reader of signs, a critic of evidence, and a guardian of epistemic humility. KEYWORDS: Clinical reasoning, Artificial intelligence, Medical epistemology, Voltaire, Sherlock Holmes, Internal medicine, Diagnostic method.