PERSON
Diane Vaughan
The Columbia sociologist whose decade-long reconstruction of the Challenger disaster gave catastrophic failure a rigorous mechanism—and whose concept of the normalization of deviance is now the most urgently needed analytical tool in every organization that has adopted AI.
Diane Vaughan spent nearly a decade reconstructing, from thousands of pages of transcripts, engineering memoranda, and flight readiness reviews, the decision chain that led to the January 28, 1986 destruction of the Space Shuttle Challenger and the deaths of seven crew members. What she found was neither incompetence nor corruption but something more unsettling: the ordinary, meticulous, institutionally rational operation of an organization that had taught itself, flight by flight, to regard as normal the conditions that would destroy it. The concept she named for this process—the
normalization of deviance—describes a four-phase mechanism: an anomaly is observed; it is assessed as acceptable given available evidence; it is reclassified as normal; it becomes the new baseline against which future anomalies are compared. At no stage does anyone make a reckless decision. At no stage does anyone intend harm. The catastrophe is the product not of one bad judgment but of twenty-four adequate ones, each of which shifted the