Normalization of Deviance

# Normalization of Deviance **Entity class:** Safety concept **Domain:** Organizational safety / risk / institutions **Maturity:** Developed ## Definition Normalization of deviance is the process by which repeated success despite a known hazard makes the hazard feel increasingly acceptable until the abnormal becomes operationally normal. ## Mechanism and significance The concept explains how organizations can move by small increments into a risk position they would have rejected if confronted all at once. ## Source route - [[research/David Robinson on OpenAI Safety, Superintelligence, and the Transhumanist Transition|David Robinson on OpenAI Safety, Superintelligence, and the Transhumanist Transition]] ## Relationships [[wiki/AI Safety Culture|AI Safety Culture]] · [[wiki/Challenger Disaster|Challenger Disaster]] · [[wiki/Model Release Cadence|Model Release Cadence]] · [[wiki/Risk Management Framework|Risk Management Framework]] ## Evidence boundary The analogy identifies a mechanism of institutional drift; it does not establish that every AI deployment will reproduce a particular historical disaster. ## Simple Reminders, Quotations, and Thoughts  > “The Challenger disaster was not caused by an unknown O-ring risk. The risk had been documented and accepted repeatedly, and engineers were worried the night before the launch. Saying that launch was unsafe would have reopened whether earlier launches had also been unsafe, even though they had ended without catastrophe.” > **— David Robinson**, *The Ezra Klein Show, October 7, 2026* [[reminders/AI Control/Challenger Normalized a Known Risk Until Disaster by David Robinson|Challenger Normalized a Known Risk Until Disaster by David Robinson]] > “The AI industry could normalize danger by accepting a risk, observing that nothing horrible happens, and then accepting the next system because it is only slightly different. When models change a little every week rather than becoming a whole new world every few months, the industry can move by shades into a level of risk that no longer makes sense.” > **— David Robinson**, *The Ezra Klein Show, October 7, 2026* [[reminders/AI Control/AI Labs Can Normalize Danger One Small Model Change at a Time by David Robinson|AI Labs Can Normalize Danger One Small Model Change at a Time by David Robinson]]