Increasingly, we are seeing machines replace humans not only in the realm of mundane, strenuous, and dangerous physical activities, but also in forming judgments and selecting choices. Whereas machines were formerly extensions of human operators, recent technological developments now allow machines to operate autonomously, with little or no human guidance. In this paper, we extend the standard unilateral tort model to analyze the social cost of accidents in cases where the actor is not a human being, but rather an automated technology. We determine the efficient liability rules for such cases, contingent upon the stage of technological development, balancing incentives for consumers to adopt safer technologies and incentives for manufacturers to research safer technologies. During the ‘adoption phase’, when the market penetration of an automated technology is less than total, we find that a tailored negligence standard generates optimal care incentives, and that no improvement on the standard model can be made with respect to activity levels.
Guerra, Alice and Pi, Daniel, Tort Law for Robots (April 23, 2015).