Autonomous Vehicles, Technological Progress, and the Scope Problem in Products Liability

Abstract

Autonomous vehicles are widely expected to save tens of thousands of lives each year by making car crashes attributable to human error – currently the overwhelming majority of fatal crashes – a thing of the past. How the legal system should attribute responsibility for the (hopefully few) crashes autonomous vehicles cause is an open and hotly debated question.

Most tort scholars approach this question by asking what liability rule is most likely to achieve the desired policy outcome: promoting the adoption of this lifesaving technology without destroying manufacturers’ incentives to optimize it. This approach has led to a wide range of proposals, many of which suggest replacing standard rules of products liability with some new system crafted specifically for autonomous vehicles and creating immunity or absolute liability or something in between.

But, I argue, the relative safety of autonomous vehicles should not be relevant in determining whether and in what ways manufacturers are held liable for their crashes. The history of products liability litigation over motor vehicle design shows that the tort system has been hesitant to indulge in such comparisons, as it generally declines both to impose liability on older, more dangerous cars simply because they lack the latest safety features and to grant immunity to newer, safer cars simply because of their superior aggregate performance. These are instances in which products liability law fails to promote efficient outcomes and instead provides redress for those who have been wronged by defective products.

Applying these ideas to the four fatalities that have so far been caused by autonomous vehicles suggests that just as conventional vehicles should not be considered defective in relying on a human driver, autonomous vehicles should not be immune when their defects cause injury.