Advances and improvements in computing power and processing have led to a clear upward progression in the degree to which autonomous vehicles can operate freely without human involvement. Advances in autonomous vehicle technology may reduce the incidence of vehicle accidents born from human error and would be a general benefit if widely used and properly regulated. However, with increases in machine agency comes the corresponding challenge of machine ethics that must keep pace with the increasing number of decisions autonomous cars need to make. In this paper, I explore and advance a view on how autonomous vehicles ought to respond in a particular tragic choice scenario under a specific set of constraints where any one person needs to die for the sake of many more. I argue that in such cases autonomous vehicles ought to randomly select who to sacrifice and that such random selection ought to be blind to particulars that set people apart from each other, including whether a potential sacrifice is a passenger or owner of the self-driving car in question.
Luvuyo Gantsho, The University of Manchester, Centre of Social Ethics and Policy
Master of Science.
Журнал “AI and Ethics”, 2022.