Now that self-driving vehicles roam the roads and have already caused injury and death,1 many talented torts scholars are reviewing the role of tort law as a regulator of this new technology and as an insurer of its victims.2 In their recent article, Automated Vehicles and Manufacturer Responsibility for Accidents: A New Legal Regime for a New Era, the formidable duo of Ken Abraham and Robert Rabin join the inquiry. Abraham and Rabin write, “The new era of automated vehicles will eventually require a new legal regime that properly fits the radical new world of auto accidents.”
Having laid down this challenge—to fit a new legal regime to a new technological era—Abraham and Rabin assess both the anticipated challenges of self-driving vehicles and the appropriate tort law responses to them. The authors expect automobiles to evolve through a number of stages, beginning with vehicles that are driver controlled, progressing to vehicles that are machine-assisted, then transitioning to vehicles that are machine-controlled but driver-assisted, and culminating in vehicles that are fully machine-controlled. Abraham and Rabin welcome this evolution. They believe that “[a]ccident rates will decline precipitously, by some estimates as much as 80-90 percent.” (P. 2.) Yet they also believe that the transition to automated technology will be “long and uneven.” Abraham and Rabin foresee periods in which varied vehicles will coexist on the roadway—much as the horse and buggy ultimately, but not immediately, gave way to the automobile.
To regulate this mixed environment, Abraham and Rabin propose plans for the short term and the long term. For the near future, when conventional vehicles predominate, Abraham and Rabin would retain the current “driver-focused” tort liability regime alongside regulation of automated technologies. As the authors explain, driver-focused negligence liability accords with the profile of current accidents, which involve frequent driver errors. Also, individualized insurance underwriting comports with a driving context in which drivers are more or less careful. Regulating automated technologies under current systems is a more difficult matter. Abraham and Rabin are pessimistic about the adequacy of regulation by the National Highway Transportation Safety Agency. Their plea is for the agency to set up ex ante performance standards to guide innovation and ex post oversight of design miscalculations. They also examine the potential for, and pitfalls of, products liability claims in cases of automation defects.
For the long run, once highly automated vehicles comprise 25% of the fleet, they would adopt a system of “’Manufacturer Enterprise Responsibility’ (MER).” (P. 17.) This system entails “auto manufacturer responsibility for all injuries arising out of the operation of HAVs (highly automated vehicles).” (P. 5.) The system would be “a manufacturer-financed, strict responsibility bodily-injury compensation system, administered by a fund created through assessments and levied on HAV manufacturers.” (P. 21.) For this system, physical harm to victims would trigger liability, without need for proof of vehicle defect.
The argument for an enterprise responsibility system stems from the authors’ belief that this system will more effectively promote safety and provide consistent and predictable compensation for injuries. With less driving by individuals, there is less reason to impose liability on individuals or underwrite insurance on an individualized basis. (P. 8.) The MER system would place liability for harm with the manufacturer, the entity that seems most capable of preventing harm. Much of the article is then focused on adopting and activating an enterprise liability system.
The strength of the Abraham and Rabin piece is its practicality. They outline the virtues of an enterprise liability system—it would be more predictable for manufacturers and more even-handed for injured parties. They draw parallels to other systems such as workers compensation and long-term disability. And they examine how the manufacturer-focused system might work in practice—through contracts with parts suppliers and caps on damages, for example. Abraham and Rabin also forthrightly address potential weaknesses of their plan. For example, enterprise liability could increase the cost of highly automated vehicles and make them more expensive to produce, even though their better safety profile would suggest encouraging rather than discouraging manufacturers to produce them. Here, Abraham and Rabin suggest government subsidies to offset the manufacturers’ increased liability costs.
The plan is thoughtful and one that policymakers would do well to engage. Enterprise liability is worthy of consideration in this area, as it has been in others. But what is perhaps most striking about the plan is how much the recommended new legal regime resembles suggestions for the old one. Indeed, the advent of highly automated vehicles seems to have little influence on the proposed legal system at all. In Professor Rabin’s 1996 article, Some Thoughts on the Ideology of Enterprise Liability, he begins with George Priest’s “core definition of enterprise liability in the context of product-related injuries.” In particular, “the twin notion that an enterprise should bear the risks of accidents it produces because (1) an enterprise has superior risk-spreading capacity compared to victims who would otherwise bear the costs of accidents, and (2) an enterprise is generally better placed to respond to the safety incentives created by liability rules than is the party suffering harm.” Abraham and Rabin’s proposal applies existing ideas of enterprise liability to a new context. This idea is sound and adaptive, but far from disruptive.
A disruptive system may not be needed. Existing theories of enterprise liability may be well suited to HAVs, like other products. The law of the HAV, like the law of the horse, may not need to be written.3 A difficulty with manufacturer enterprise liability seems to be a difficulty with enterprise liability in general. HAVs seem like many products—one of multiple causes of accidents. Current auto accidents are caused by a combination of driver failure, vehicle design, and driving context, particularly road design. After the shift to HAVs, even if driver error can be completely eliminated from the accident equation, driving context will likely remain as important, if not more important a factor in traffic accidents.
Even after automation, ordinary contextual elements like rain and snow can play a role in crashes. After automation, newer contextual elements may play a role as well. Through vehicle-to-vehicle communication, the function of nearby vehicles and roadway networks may become a particularly important determinant of accidents. Network design, and even events like theft of valuable metals from smart roadways can interfere with car safety. If highly automated vehicles are one factor among others in roadway crashes, in this context, as in others, scholars will need to make the case for the efficacy and fairness of enterprise liability more broadly.4 In the HAV context, as in other enterprise liability contexts, scholars will have to wrestle with the difficult question of when an accident should be attributed to an automated vehicle rather than weather, road design, or other factors.
Another possibility, is that this transformative technology really will warrant a completely new legal regime. As we enter an era of HAV’s and more sophisticated technology, perhaps those technologies themselves might be designed to better regulate accidents or compensate injured parties. HAVs could be required to record and transmit automated records of accident or near-accident data and update federal regulatory performance standards directly or pay fines per design miscalculation (running a red light, for example), rather than per injury. Imagine disputes settled through crowdsourcing.
It may be that the more things change the more they stay the same, and that new technologies can be regulated with existing ideas. But it may also be that we will not fully envision the new legal regime that best fits radical new technologies until those technologies are not just roaming the streets, but embedded in the streets themselves.
- Daisuke Wakabayashi, Self-Driving Uber Car Kills Pedestrian in Arizona, Where Robots Roam, N.Y. Times (March 19, 2018). [↩]
- See, e.g., Mark Geistfeld, The Regulatory Sweet Spot for Autonomous Vehicles, NYU Public Law & Legal Theory Working Paper No. 18-29 (May 2018). [↩]
- Frank H. Easterbrook, Cyberspace and the Law of the Horse, 1996 U. Chi. Legal F. 207, 207–08 (1996) (“Dean Casper’s remark [that he refused to offer a course on the law of the horse] had a second meaning–that the best way to learn the law applicable to specialized endeavors is to study general rules”). [↩]
- See, e.g., Gregory C. Keating, Products Liability as Enterprise Liability, 10 J. Tort L. 1 (2017). [↩]