Document Type


Publication Date

Fall 2018


Self-driving vehicles, Artificial intelligence, Products liability


Consumer Protection Law


The trolley problem is a well-known thought experiment in moral philosophy, used to explore issues such as rights, deontological reasons, and intention and the doctrine of double effect. Recently it has featured prominently in popular discussions of decision making by autonomous vehicle systems. For example, a Mercedes-Benz executive stated that, if faced with the choice between running over a child that had unexpectedly darted into the road and steering suddenly, causing a rollover accident that would kill the driver, an automated Mercedes would opt to kill the child. This paper considers not the ethical issues raised by such dilemmas, but the liability of vehicle manufacturers for injuries that foreseeably result from the design of autonomous systems. Some of the recent commentary on the liability of autonomous vehicle manufacturers suggests unfamiliarity with modern products liability law, particularly the design-defect standard in the Third Restatement of Torts. A superficial understanding of products liability principles – for example, believing it is a regime of strict liability in any meaningful sense – can lead to serious errors in the application of this area of law to autonomous vehicles. It is also a mistake to believe that the economic approach to negligence liability, as developed by Posner and Calabresi, accurately characterizes modern products liability principles. Under the Third Restatement approach, a court or jury will consider whether a product embodies a reasonable balance of safety and utility, and “reasonable” can be interpreted in accordance with ordinary community ethical standards. Thus, some of the issues that are central to resolving trolley problems in moral philosophy may actually recur in design-defect litigation.


This paper is a contribution to a symposium on Artificial Intelligence and Real World Ethics, at California Western School of Law.