Aggregating Small Risks of Serious Harms

Tomi Francis (Global Priorities Institute, University of Oxford)

GPI Working Paper No. 21-2024

According to Partial Aggregation, a serious harm can be outweighed by a large number of somewhat less serious harms, but can outweigh any number of trivial harms. In this paper, I address the question of how we should extend Partial Aggregation to cases of risk, and especially to cases involving small risks of serious harms. I argue that, contrary to the most popular versions of the ex ante and ex post views, we should sometimes prevent a small risk that a large number of people will suffer serious harms rather than prevent a small number of people from certainly suffering the same harms. Along the way, I object to the ex ante view on the grounds that it gives an implausible degree of priority to preventing identified over statistical harms, and to the ex post view on the grounds that it fails to respect the separateness of persons. An insight about the nature of claims emerges from these arguments: there are three conceptually distinct senses in which a person’s claim can be said to have a certain degree of strength. I make use of the distinction between these three senses in which a claim can be said to have strength in order to set out a new, more plausible, view about the aggregation of people’s claims under risk.

Other working papers

Imperfect Recall and AI Delegation – Eric Olav Chen (Global Priorities Institute, University of Oxford), Alexis Ghersengorin (Global Priorities Institute, University of Oxford) and Sami Petersen (Department of Economics, University of Oxford)

A principal wants to deploy an artificial intelligence (AI) system to perform some task. But the AI may be misaligned and aim to pursue a conflicting objective. The principal cannot restrict its options or deliver punishments. Instead, the principal is endowed with the ability to impose imperfect recall on the agent. The principal can then simulate the task and obscure whether it is real or part of a test. This allows the principal to screen misaligned AIs during testing and discipline their behaviour in deployment. By increasing the…