Ethical considerations and moral implications of autonomous vehicles and unavoidable collisions

Abstract While it is widely agreed that automated and autonomous vehicles may provide safety benefits over vehicles with lower level or no automation, due to other road users there will still likely be situations where a collision is unavoidable. What should a vehicle that is operating autonomously do when it has no choice but to have a collision? And who should decide which vehicle manoeuvre is the most acceptable? These situations create moral dilemmas requiring consideration of the most acceptable and moral action of the vehicle. In this paper we explore current research in this domain and work towards enabling ethical solutions. We identify current experimental work (practical studies rather than theoretical studies) on this problem often contains fundamental flaws due to the lack of real-world validity within the studied scenarios. We argue that morality is highly context dependent and that participants need to be more engaged in the choices they are claiming to make. Suggestions for future work include virtual reality or simulation methodologies which promote immersivity to ensure procedural validity whilst retaining safety. We also identify current guidelines contradict public viewpoint and argue public attitude needs to be better understood to give autonomous and automated vehicle manufacturers confidence in their design.