Abstract
This article presents a study of how automobile risks are changing over time as the level of automation shifts from the traditional to semi-autonomous to fully autonomous car. The study involves the concept of comparative advantage between the human driver and the driverless car to determine how the roles between the two have been changing and, in turn, how the change affects attendant risks. To develop a risk landscape across changes in levels of automation, the roles of human driver and driverless car are matched with the five levels of automation used to describe the phases of development of the driverless car. We conclude that the costs associated with fatal accidents will decrease sharply over time as the driverless cars are deployed on the road. And we offer implications for policy, regulation, and insurance based on our analysis of the changing risk landscape.