CONCEPT
Risk versus Uncertainty
Jasanoff's foundational distinction: risk covers outcomes specifiable in advance with assignable probabilities; uncertainty covers emergent consequences no model anticipates.
Risk and uncertainty are not synonyms but categorically different epistemic conditions requiring different governance approaches. Risk refers to outcomes that can be specified in advance and assigned probabilities — bridge collapses, pharmaceutical side effects, algorithmic discrimination. Risk is the domain of prediction and expert assessment. Uncertainty refers to outcomes that cannot be specified because they arise from interactions between systems whose combined behavior is emergent. The most important consequences of AI — what happens to professional identity when expertise is commoditized, what happens to children's cognitive development when intellectual effort can be bypassed, what happens to democratic culture when persuasive content costs nothing to produce — are uncertain in this precise sense. They depend on interactions between technology and social order that have never existed before, producing outcomes no participant can predict. Governance institutions designed for risk management fail under uncertainty because they rely on prediction, and the consequences that matter most cannot be predicted.
In The You On AI Field Guide
Jasanoff's distinction builds on Frank Knight's 1921 Risk, Uncertainty, and Profit