Abstract

The basic question for risk assessment is not "What are the risks?" but "How safe is safe enough?" Its ambitious goal is to make risk management a scientific enterprise. In order to succeed, not only must risks be quantified but also the many kinds of costs and benefits associated with technology and its control must be quantified and we must find a common metric for comparing these different factors. The risks of risk assessment include the possibility of distorting values in the quest for general comparability and also the problem of responding rationally to different kinds of uncertainty. The problems discussed include comparing novel events to statistically frequent events, comparing different distributions of risk, and quantifying the value of saving human lives.