Mark Raffman, Goodwin Procter partner (Photo: Courtesy photo)

Introduction

For the last half century, autonomous vehicles have been more in the realm of science fiction than reality. However, technological advancements in self-driving vehicles are slowly transforming this fiction into reality. With safety features like lane control and emergency braking quickly becoming part of our daily lives, state governments must prepare for the inevitable shift from nonautomated to fully automated vehicles. See John Gardella, “Changing Gears: Shifting Autonomous Vehicle Liability,” Claims 23, 23-25 (May 2019) (Gardella).

In July 2019, Tesla announced that it planned to roll out an “over-the-air” software update converting hundreds of thousands of its vehicles into largely autonomous vehicles by the end of the year, prompting concerns of preparedness among policymakers, see Faiz Siddiqui, “Tesla floats fully self-driving cars as soon as this year. Many are worried about what that will unleash,” The Washington Post (July 17, 2019). With fully automated cars projected to reach consumers by 2030, states must anticipate these changes now, and how they will impact citizens and infrastructure going forward.

Despite concerns about the rapid growth of autonomous vehicle technology, the development of remotely-operated and semi-autonomous vehicles will benefit the United States economically and environmentally. The development of autonomous vehicles will improve vehicle safety, increase energy efficiency and reduce traffic on our roads, see “Preparing for the Future of Transportation: Automated Vehicles 3.0,” U.S. Department of Transportation 1, 16 (Oct. 2018) (Automated Vehicles 3.0); Catherine Ross & Subhrajit Guhathakurta, “Autonomous Vehicles and Energy Impacts: A Scenario Analysis,” 143 Energy Procedia 1, 47 – 52 (Dec. 2017).

With such benefits in mind, federal and state governments should be doing what they can to further the development of this technology (especially as it pertains to dynamic ridesharing), while yielding a consistent regulatory landscape that protects the public interest.

Since 2012, at least 41 states and the District of Columbia have considered legislation related to autonomous vehicles and remotely operated vehicles. See “Autonomous Vehicles: Self-Driving Vehicles Enacted Legislation,” National Conference of State Legislatures (March 19, 2019) (NCSL). While these statutes represent a good first attempt to address the rise of this new technology, many fail to go in depth regarding potential issues surrounding their future. We recommend that states focus on adopting legislation in four key areas to promote the secure operation and financial accountability of autonomous vehicles in the years to come: insurance requirements; provisions for licensing or testing; driver impairment; and accidents and liability plans.

Background

Policy proposals in the area of autonomous and remotely operated vehicles require a basic understanding of the levels of autonomy at which these vehicles may operate. As of 2019, the Society for Automotive Engineers International (SAE) has identified six key stages of vehicle automation. Recognized as the applicable standards by the U.S. Department of Transportation (DOT), see “Automated Driving Systems 2.0: A Vision for Safety,” National Highway Traffic Safety Administration (Sept. 12, 2017), these automation stages help to guide the development of autonomous vehicle regulation moving forward:

• Stage 0 (No Automation): Driver performs all driving tasks;
• Stage 1 (Driver Assistance): Driver controls the vehicle, but some assistance features may be available;
• Stage 2 (Partial Automation): Some automated features exist, but driver must remain engaged in driving tasks and monitor environment at all times;
• Stage 3 (Conditional Automation): Driver is needed, but not required, to monitor the environment and must be ready to take control of the vehicle at all times;
• Stage 4 (High Automation): Vehicle can perform all driving functions under certain conditions. Driver may have the option to control the vehicle; and
• Stage 5 (Full Automation): Vehicle can perform all driving functions under certain conditions. Driver may have the option to control the vehicle.

While these stages encapsulate the future scope of autonomous and remotely operated vehicles, many of the later stages (Stages 4 and 5) are still years away. At present, the most advanced autonomous vehicle currently available on the market is only a Stage 2 partially autonomous vehicle.

Insurance Requirement

As the use of autonomous and remotely operated vehicles increases, state insurance monitors will need to adapt accordingly. While a clear majority of states have yet to address insurance requirements for autonomous vehicles, a handful of states (New York, Nevada, Connecticut, Florida, Georgia, and Nebraska), have implemented mandatory insurance requirements for both autonomous vehicle testing and road use.

One of the measures, Georgia SB 219, see GA. CODE ANN. Section 40-8-11(a)(4)(A) (2017), may serve as a model for other states contemplating commercial insurance requirements for autonomous and remotely-operated vehicles. Under the amended statute, drivers of autonomous vehicles are required to obtain indemnity, liability, and self-insurance equivalent to 250% of what is required under Georgia’s commercial indemnity and liability insurance statute, see GA. CODE. ANN. Section 40-1-166, (a minimum of $300,000). While this may seem punitively high, the 250% requirement is set to expire on Dec. 31, 2019—effectively mandating high insurance costs for autonomous vehicles in their infancy.

At present, we are most likely to see the development of remotely operated or autonomous vehicle technology in a commercial application, likely in sectors involving long-haul trucking and dynamic ridesharing. As companies in these sectors may be expected to have both the financial resources to be able to pay high insurance rates and the incentive to continue developing autonomous vehicle technology—at least by the time their vehicles are ready to operate on public roadways—these early insurance regulations (like Georgia SB 219) will help provide financial protection from casualties without interfering with technological advancement. Basing state insurance rates for autonomous and remotely-operated vehicles on existing rate requirements for commercial vehicles will provide states with a uniform way to address the increased usage of autonomous vehicles while balancing the continued presence of non-autonomous vehicles on the same roadways.

The rise of autonomous and remotely operated vehicles will ultimately change the structure by which liability will be determined in vehicle accidents, necessitating some modifications to the way in which insurance addresses such accidents. Although autonomous vehicles will likely diminish the overall incidence of vehicle casualties by eliminating aspects of “human error,” future accidents and claims may be caused by equipment and software errors in the vehicles—likely yielding grounds for liability for manufacturers, either in addition to or instead of vehicle operators. Whereas the process for determining liability for today’s motor vehicle accidents is a familiar process, with insurance companies generally able to adjust and settle claims between and among drivers, the process for adjudicating liability in future autonomous vehicle accidents will involve new actors such as vehicle manufacturers, component manufacturers, and local governments, potentially creating new and complex procedural and substantive twists. The complexity of addressing liability for autonomous vehicle accidents could be reduced if product liability insurance policies for autonomous vehicles were to include all entities that participate in the manufacture of the vehicle (e.g., component suppliers, manufacturers, dealers), thus limiting the complexity of adjudications to provide recompense to injured parties. Separately, state and municipal governments that install sensors or other equipment to facilitate the operation of autonomous vehicles on public roadways will need to ensure that their liability insurance policies include coverage for failure to install, maintain, or repair such equipment, as such government inaction could result in malfunctions and accidents in autonomous vehicles.

Provisions for Licensing or Testing

State governments should also be prepared to establish licensing or testing requirements for autonomous and remotely-operated vehicles. While states have thus far been silent on licensing requirements for fully autonomous cars (likely due to the fact that this technology has yet to progress to a Stage 4 or 5, where human intervention is unnecessary), certain states have begun considering the licensing requirements for truck “platoons”—i.e. “a group of individual commercial trucks traveling in a unified manner at electronically coordinated speeds at following distances that are closer than would be reasonable and prudent without the electronic coordination.”

One model that may serve as a starting point for state governments considering establishing licensing requirements for remotely operated vehicles or platoons can be found in a measure recently passed by the Kentucky legislature. See KY. REV. STAT. ANN. Section 281.010 (West 2018). Under Kentucky Revised Statute Section 281.10, in order to operate a platoon, a driver must hold a valid commercial driver’s license and be present behind the wheel of each vehicle in the platoon. Although the latter requirement may ultimately prove problematic in the case of remotely-operated vehicles—i.e., vehicles controlled from a distance—for the time being it provides some assurance that human intervention is available when needed. Further, the requirement that an operator hold a valid commercial driver’s license is a ready way to promote safe operation of platoons. The requirements necessary for a commercial driver permit (which include a skills test, road test, and vehicle inspection controls exam) can easily be adapted by state officials to meet the evolving needs of autonomous and remotely operated vehicle technology.

Rules for Driver Impairment

Another area virtually unaddressed by the states to date is the potential impairment of individuals in the autonomous vehicle context. Here, there are two sets of concerns: how to determine impairment in the case of individuals controlling remotely-operated vehicles; and what role impairment plays in the case of individuals who are “operating” fully autonomous vehicles. In crafting new laws concerning driver impairment, lawmakers should be careful to set out distinct rules for autonomous vehicles and remotely operated vehicles respectively, as both technologies carry with them different problems and intricacies.

Under no circumstances should the driver of a remotely operated vehicle operate such a vehicle while under the influence of alcohol or drugs. Though the driver of a remotely-operated vehicle may be hundreds of miles away from the vehicle at the time of operation, the driver is still acutely in charge of the vehicle and its movements. Under most state laws, the legal limit for the blood alcohol concentration (BAC) for drivers is 0.08%, while the BAC for commercial drivers is 0.04%. See DUI and DWI, Department of Motor Vehicles (July 8, 2019). However, according to a study conducted by the National Institute of Health, a BAC of 0.03% or higher will have an impact on a driver’s perception and reaction times. See National Institute on Alcohol Abuse and Alcoholism No. 25 PH 351, National Institute on Alcohol Abuse and Alcoholism (July 1994). As the study explained, “low to moderate BACs (0.03 to 0.05 percent) interfere with voluntary eye movements, impairing the eye’s ability to rapidly track a moving target,” and “significant impairment in steering ability may begin as low as approximately 0.035% BAC and rises as BAC increases.”

Impacts of alcohol on eye movement and steering ability have a direct impact on operators of remotely operated vehicles, making them at least as susceptible to an impairment-related collision as a driver of a nonautomated vehicle. Given the fact that remotely operated drivers are removed from the vehicle environment (operating a vehicle from the comfort of a desk instead of within the vehicle), they may be even more susceptible to impairment-related errors than drivers in non-automated vehicles. As a result, states should impose no more lenient standards for vehicle impairment of non-automated vehicles on drivers of remotely-operated vehicles. Companies with fleets of remotely-operated vehicles should also consider instilling a zero-tolerance policy for alcohol or drug consumption on company grounds, accompanied by frequent mandatory employee drug tests.

The question, then, is how state authorities can determine impairment in the case of remotely operated vehicles. Whereas the officer on the scene can easily administer a field-sobriety test to the driver of an ordinary vehicle who is suspected of impairment, the driver of a remotely operated vehicle is, by definition, physically located away from the scene of the accident, likely even in a different state. Yet the need to determine sobriety in the event of an accident involving a remotely operated vehicle is no less acute. States can address this issue by requiring, as a condition for the operation of remotely-operated vehicles within their borders, that companies operating such vehicles consent to the testing of their employees involved in a casualty and undertake specific practical measures to ensure that such testing will be conducted on request and the results shared with state authorities.

With respect to operators of autonomous vehicles, a slightly different calculus is in order. For vehicles that are partially or fully dependent on the driver, of course, the rules need not and should not change. But in the case of fully autonomous Stage 4 or 5 vehicles, the occupant is arguably reduced to the status of a passenger. Therefore, some international regulators have recommended an exemption for impaired driving when a vehicle is fully-autonomous. As the Australia National Transport Commission explained: the application of an exemption is clear cut for dedicated automated vehicles, which are not designed for a human driver. The occupants will always be passengers … any exemptions should not apply to the fallback-ready user of a vehicle with conditional automation. A fallback ready user is required to be receptive to requests to intervene to system failures and must take over the dynamic driving task if the ADS cannot perform it. See “Changing Driving Laws to Support Automated Vehicles: Discussion Paper,” Australia National Transport Commission 1, 70 (Oct. 2017).

But the day when fully autonomous vehicles can safely convey impaired occupants home after a night of carousing is still in the future. For states contemplating a possible exemption for impaired drivers of autonomous vehicles, they should be clear that the exemption applies only to those Stage 4 or 5 autonomous vehicles. As noted above, the most advanced autonomous vehicle currently available on the market is only a Stage 2 vehicle—a partially autonomous vehicle under the control of the driver. Until the day that Stage 4 or 5 autonomous vehicles become commercially available, it is imperative that states make clear that operating an autonomous vehicle under the influence of drugs or alcohol is still criminally liable behavior.

Accidents and Liability Plans

Finally, when preparing for the arrival of remotely-operated and autonomous vehicles, state legislatures should consider developing regulations to provide safe harbor for validated operating algorithms. A validated operating algorithm is a formula agreed upon by car manufacturers and state or federal governments which outlines a series of design choices that should be made by the autonomous vehicle in the event of a potential accident. When an autonomous vehicle operates in compliance with its validated operating algorithm (e.g., swerving into an opposing lane to avoid a sure collision) the vehicle, its operator, and the manufacturer should have safe harbor from potential criminal liability. Developing a plan for potential collisions involving autonomous vehicles is one of the top recommendations for manufacturers made recently by the DOT in Automated Driving Systems: A Vision for Safety 2.0. In the release, the DOT stated that “entities are encouraged to have a document process for assessment, testing, and validation of their crash avoidance capabilities and design choices.” See NCSL. DOT recommends that this document include pre-crash scenarios related to lane change/merging collisions, head-on collisions, control loss, crossing path crashes and low-speed situations as parking maneuvers.

When an autonomous vehicle operates in conformity with its validated operating algorithm, states should also consider excusing the operator or manufacturer from liability for exemplary damages. The advent of algorithms that yield a predetermined “choice” about what a vehicle will do in the case of a casualty situation may tempt juries to penalize the entity whose “choice” resulted in an injury to a specific person. Yet overall these algorithms—especially when subjected to regulatory scrutiny and approval—will yield safer outcomes overall. By ensuring that a vehicle’s driver or manufacturer will not face additional liability over pre-approved programming, states will encourage the development of safe remotely operated and autonomous vehicles and the rational computation of accident liability.

Conclusion

Once considered futuristic technology, autonomous and remotely-operated vehicles are becoming part of our daily lives. Beginning with safety measures like lane control and emergency braking, modern society is adapting to increasingly automated vehicles. While Stage 4 or 5 autonomous vehicles are a decade away, state legislatures should begin planning for and adapting state law to meet this developing technology head-on. By crafting laws and regulations concerning insurance requirements, licensing and testing, driver impairment, and liability plans for autonomous and remotely-operated vehicles, states will be prepared for the secure use and financial accountability of these future technological developments—no matter how “science-fiction” they seem now.

—Katherine Kissinger, a recent summer associate with the firm, contributed to this article.

Mark Raffman is a partner in law firm Goodwin’s Washington, D.C. office where he concentrates his practice on complex products liability and consumer products litigation and advises on regulatory compliance audits and legislation as well as transactions posing products liability risks. He has 30 years’ experience defending complex litigation matters in state and federal courts nationwide, is a member of the Defense Research Institute and has lectured and published on tort liability litigation developments, most recently on products liability risks.