People make suboptimal decisions about existential risks

Abstract

Allocating resources to maximize the probability that humanity survives a set of existential risks has a different structure from many decision problems, as the objective is the product of the probabilities of desired outcomes rather than the sum. We derive the optimal solution to this problem and use this solution to evaluate the choices that people make when presented with decisions that have this multiplicative structure. Our participants (total N=2,072) are appropriately sensitive to how responsive a risk is to investment, but are conservative in their decisions and do not allocate enough resources to risks with lower probability of survival. This pattern persists even with alternative framings that emphasize survival probabilities. Our results highlight a systematic flaw in people's intuitions about how to respond to existential risks, and suggest that people may have particular difficulty with decisions that involve multiplicative objectives.