Iatrogenesis

In The You On AI Encyclopedia

The three-level structure is Illich's most transferable analytical tool. What medicine does to health, other institutions do to the goods they claim to provide. Applied to AI, the structure maps with painful precision. Clinical: the direct errors AI produces—fluent fabrication, confident wrongness, hallucinated citations that pass superficial review and fail only under scrutiny that the tool's fluency discourages. Social: the dependency created when knowledge workers cannot perform without AI mediation, when expertise is redefined as AI-augmented expertise, when organizational expectations recalibrate around augmented output and the unaugmented worker falls below the threshold of acceptable performance. Cultural: the loss of the human capacity to think, struggle, and produce understanding through cognitive effort—capacities AI intervention progressively replaces with model-generated output, until the population can no longer engage with intellectual difficulty without institutional mediation through the tool.

The mapping is not analogical. It is structural. The same institutional logic operates in both domains. An institution designed to serve a need grows. As it grows, it professionalizes. As it professionalizes, it acquires authority to define the need. As it defines the need, it delegitimizes alternatives. Dependency increases. Autonomous capacity atrophies. The institution becomes more necessary. The cycle feeds itself.

What makes cognitive iatrogenesis distinctive, and potentially more consequential than its medical predecessor, is that it operates on the instrument of detection itself. Medical iatrogenesis can be measured because the patient retains the capacity to report symptoms, and because external observers can examine outcomes. Cognitive iatrogenesis is harder to measure because the degraded capacity is the capacity by which degradation would be noticed. The debugger who can no longer debug does not notice, because the tool debugs for her. The writer who can no longer structure an argument does not notice, because the tool structures arguments beautifully.

Illich's response was institutional limits—political constraints on medical system growth, calibrated to the point at which expansion began to generate iatrogenic harm at scale. The analogous response for AI would be limits calibrated to the point at which deployment begins to generate cognitive harm at scale. But cognitive harm cannot be measured with the instruments currently available, and the institutions with the authority to impose limits have no incentive to develop measurement systems that would justify constraint.

Origin

The term existed in clinical literature before Illich, but Medical Nemesis (1975) expanded it into a systematic analytical framework with three distinct levels. The expansion drew on WHO epidemiological data, historical studies of medicalization, and Illich's direct engagement with health-policy debates in the late 1960s and early 1970s.

The framework has been absorbed into mainstream medical ethics, patient-advocacy movements, and increasingly into technology ethics, where it provides the structural template for analyzing institution-induced harm.

Key Ideas

Clinical level. Direct harm of the intervention—the errors, fabrications, and failures that survive delivery to the user.

Social level. Dependency created when populations cannot perform the activity without institutional mediation.

Cultural level. The loss of autonomous capacities the institution replaces—the deepest and most irreversible form of harm.

Compounding structure. Each level generates more of itself; harm accumulates across levels until the institution becomes structurally necessary.

AI mapping. Fluent fabrication, workplace dependency, and cognitive atrophy correspond precisely to Illich's three levels of medical iatrogenesis.