Law and Economics of Artificial Intelligence : Optimal Liability Rules for Accident Losses Caused by Fully Autonomous Vehicles

We consider the optimal liability rule in the case of accidents by fully autonomous vehicles. We show that when due activity is not enforceable, it is socially optimal to apply the strict liability rule to the human operator who determines the activity level and to apply the negligence rule to the manufacturer and the victim who choose care levels under contributory or comparative negligence in the unilateral activity case. Under the joint and several liability rule, both the manufacturer and the victim take due care if the liability share of the manufacturer can be regulated high enough, and the human operator who must take all the remaining risks chooses the socially optimal activity level that maximizes the social net benefit. However, if it is possible to enforce due activity, the opposite liability rule will be optimal. Under the rule, the human operator takes efficient activity to comply with the activity standard and the manufacturer takes efficient care to meet the care standard, and the victim bears residual liability, which clearly induces him to take efficient care. This liability rule can achieve social optimum, even in the bilateral activity case in which both the human operator and the victim take activity. Our results are distinguished from previous findings that social optimum can be attained by using the public sanction of paying a fine to the state