Self-Driving Cars, Trolley Problems, and the Value of Human Life: An Argument Against Abstracting Human Characteristics

Abstract The rise of self-driving cars raises numerous ethical conundrums, none has attracted so much public attention as the question of how to programme AVs in crash scenarios. How does a car respond when difficult, life-and-death choices are to be made? The most popular approach to answering this question is to employ trolley problems (trolleyology). Trolleyology, employed within the context of AVs, pits one human life against another on the basis of their distinctive characteristics: old vs young, single vs group membership, innocent vs guilty, etc. Doing so attempts to create value hierarchies for human particularity. This article argues for the failure of this approach by considering the distinctive manner in which trolley problems are employed within crash scenarios involving AVs. The article argues for the evaluative error of establishing human value on human particularity through a process of abstract individualisation. Properly speaking, individual humans are valued not on the basis of their particularities, but on the basis of their innate unsubstitutability.