Which World Gets Saved

Abstract

Arguments for prioritizing existential risk (x-risk) reduction often cite the immense potential value of the future. While common counterarguments suggest the future could be net-negative or that improving its trajectory might be a better use of resources, another key consideration is often overlooked. The expected value of the future, conditional on the successful aversion of a specific x-risk, may differ significantly from the unconditional expected value. This is because averting a risk provides evidence about the nature of the world. For instance, if an effort to prevent a nuclear war succeeds, this implies we inhabit a world where such a conflict was a serious possibility, potentially one with an inherent disposition towards violence that lowers its long-term prospects. This “which world gets saved” problem complicates the simple maxim to “minimize existential risk,” suggesting a need to differentiate between risk-reduction efforts, as some may disproportionately save worlds with lower future value than others. – AI-generated abstract.