Self-Driving Vehicles, Autonomy and Justice

Self-driving vehicles, as it is widely believed, will need to be equipped with socalled ethics settings, reducing the fatalities in unavoidable crash situations. These ethics settings, as it is also widely believed, should be mandatory and impartially distribute fatalities between vehicle passengers and pedestrians, preferring the side where more lives can be saved. This paper criticizes both of these beliefs, focusing on the tension mandatory ethics settings are about to generate between values of autonomy and justice. The central argument is that mandatory ethics settings, as long as they rely only on the number of lives to be saved or sacrificed, fail to respect one elementary principle of justice (>> equals should be treated equally and unequals unequally <<) by disregarding some significant differences between self-driving vehicle passengers and pedestrians. After rejecting a possible reply to this argument (that mandatory ethics settings could also rely on certain qualitative features ofparticipants in unavoidable crash situations), it is concluded that no generally acceptable normative solution to the ethics settings problem is likely and that technological approaches to reducing fatalities in traffic with self-driving vehicles should be preferred.