PERSON
Charles Perrow
The Yale sociologist who proved that the most dangerous feature of complex technologies is not malfunction but normal operation—the catastrophes that tightly coupled, interactively complex systems produce not in spite of their design but because of it.
Charles Perrow is the theorist of the inevitable accident. For four decades he studied how high-risk organizations fail, and the conclusion he reached was more disturbing than any diagnosis of incompetence or negligence could be: in systems that combine
interactive complexity with
tight coupling, catastrophic failure is not an aberration but a structural certainty. The operators who made things worse at Three Mile Island were not fools; they were skilled professionals acting exactly as their training prescribed, within a system whose architecture guaranteed that normal skill would be insufficient. Perrow published
Normal Accidents: Living with High-Risk Technologies in 1984, mapping nuclear plants, chemical facilities, aviation, and financial markets onto a two-dimensional matrix and proving that certain systems, by virtue of where they sit on that matrix, will produce
normal accidents—accidents that arise from the normal functioning of the system. He never wrote a word about AI. But the researchers who built the AI safety field adopted his framework