Self-driving cars and the bumpy road ahead | The American Association For Justice Archive

Self-driving cars and the bumpy road ahead

Trial Magazine logo, registered



February 2016 - Christopher B. Dolan


As autonomous vehicles become a reality, they have great potential to reduce crashes and protect consumers from injuries on the road. But there are many unanswered questions about liability and regulation, and automakers are trying to draft the playbook.

The high-tech crowd and the National Highway Traffic Safety Administration (NHTSA) think so. According to NHTSA, America’s drivers crashed more than 5.3 million times, injured 2.2 million people, and killed another 32,367 in 2011 alone.1 A 2005–2007 NHTSA study revealed that most crashes were the result of driver error: about 41 percent were recognition errors (inattention, internal and external distractions); 34 percent were decision errors (driving aggressively or too fast); and 10 percent were performance errors (overcompensation, improper directional control). About 18 percent of the at-fault drivers engaged in at least one non-driving activity; the most frequent was conversation, either with other passengers or on a cell phone.2

Autonomous vehicles (AVs), however, don’t make human mistakes, and there is strong evidence that AV technology may dramatically alter the collision matrix. As drivers increasingly rely on this technology, injuries will decline and lives will be saved. It’s estimated that AV technology will likely lead to a 90 percent reduction in crashes, saving nearly 30,000 lives and preventing millions of injuries annually.3 But it doesn’t stop there: In a self-driving future, traffic jams would be a thing of the past and every stoplight would be green. And AVs will cut back gasoline consumption from 1.9 billion gallons to 190 million gallons each year.4

NHTSA classifies AV technology in five levels:5

  • No automation (level 0): The driver is in complete and sole command of the primary vehicle controls.
  • Function-specific automation (level 1): Automation of one or more specific control functions, such as electronic stability control or antilock brakes.
  • Combined function automation (level 2): Automation of at least two primary control functions, such as adaptive cruise control and lane centering.
  • Limited self-driving automation (level 3): Drivers can cede full control of all safety-critical functions so long as they are available for occasional control with sufficient warning time.
  • Full self-driving automation (level 4): The vehicle performs all safety-critical driving functions and monitors roadway conditions for an entire trip, occupied by people or not.

How AVs Work

Existing AV technology relies on computer-mapping systems that generate a detailed 3-D map of the environment. When combined with high-resolution maps, radar, and cameras, these systems allow an AV to drive itself while avoiding obstacles and respecting traffic laws. Although the technology has significantly advanced, these systems are not yet capable of full AV operation.

They do not perform well under changing conditions, such as con­struc­tion zones, posted reduced speeds, or police-directed traffic flow. Development of 3-D maps is also lacking; in 2014, Google reportedly completed 2,000 miles of 3-D mapping around its Silicon Valley headquarters—just .05 percent of the 4 million miles of roads in the United States.6

Vehicle-to-Vehicle. One model for operating AVs is through vehicle-to-vehicle communication (V2V), where onboard short-range radio devices transmit data—on vehicle speed, direction, braking, and more—between vehicles. NHTSA has declared regulatory jurisdiction over V2V technology pursuant to the 1966 National Traffic and Motor Vehicle Safety Act,7 with an initiative called Intelligent Transportation Systems (ITS). In an August 2014 report, NHTSA outlined the government’s vision of the future.8 Make no mistake, the automated vehicle is not an “if” but a “when.”

One V2V application, Intersection Movement Assistance (IMA), warns drivers when it is not safe to enter an intersection because of a high probability of a collision.9 V2V technology offers forward-collision avoidance, blind-spot warnings, and lane assist features similar to those already found in many cars today—with the exception that V2V systems can use those features over greater distances, allowing cars to “see around and through” traffic for danger ahead. NHTSA estimates that a fully mature V2V system could eliminate as many as 81 percent of all light-vehicle crashes and 81 percent of all heavy-truck crashes annually.10

Vehicle-to-Infrastructure. NHTSA is also investigating a vehicle-to-­infrastructure (V2I) system, where vehicles wirelessly “talk” to communications infrastructure rather than to each other. The infrastructure coordinates the movements of multiple vehicles over its grid. V2I technology currently focuses on warning motorists about features such as red lights, stop signs, bad weather ­conditions, approaching curves, and construction zones so they can alter their behavior.

The government is funding both V2I and V2V development and testing throughout the country—the Department of Transportation (DOT) is supporting projects that analyze these communication systems in Arizona, California, Florida, Michigan, Minnesota, New York, and Virginia, and Congress has allocated $100 million annually to fund NHTSA’s research and development of V2I and V2V programming and infrastructure.11 Proponents of the V2I model want to create nationally interoperable systems.12

Hardware and software algorithms. V2I and V2V systems are only part of the emerging technology. Automakers, software companies, and component manufacturers are moving forward with ­vehicle-based AV systems that use interconnected onboard technologies such as radar, lidar, cameras, and software to allow full, independent navigation.

Manufacturers such as Volvo, Audi, Mercedes, BMW, Tesla, and GM have well-developed AV programs. And tech companies Google and Apple are also getting into the game. Google’s autonomous vehicles have driven over 1.7 million miles in autonomous mode (with a driver ready to take over in an emergency) with only 12 collisions, all attributed to driver error.13

Questions of Liability

In a collision involving an AV, who is responsible? Is it the passenger, owner, manufacturer—or all three? The calculus of duty is uncertain, and this issue lies at the heart of policy, legislative, and regulatory discussions.

Several states, including California and Nevada, have determined that the AV “operator” is liable for any injury or damage caused by the vehicle. The “operator” is the person who is either behind the controls available for manual use, or, in the absence of manual ­controls, the one who “causes the technology to engage.”14 Other states permitting AVs are silent on the issue, for the most part.15 Under Nevada and Florida law, a vehicle manufacturer is immune from any liability for aftermarket adaptation that converts a stock vehicle into an AV.16 ­Currently, California, Nevada, and Florida­ mandate that an AV ­manufacturer or developer carry $5 ­million in liability insurance.17 The unanswered liability issues could prevent driverless cars from gaining widespread consumer use18—delaying safety advancements, traffic-congestion relief, and mobility for seniors and people with disabilities.19

Some tort reformers claim that the civil justice system’s protections should not apply to AV technology. For example, one suggested approach for car manufacturers to avoid strict products liability is to reclassify the AV system as a service rather than a product. Then, contract law can limit liability, better control the way products are used, and create profits for the manufacturer.20

A call for national no-fault. RAND Corporation recently published a guide anticipating that a liability shift from the driver to the manufacturer will occur with increasing AV use—the manufacturer will be the likely target when things go wrong.21 The guide states that this shift “may make no-fault automobile insurance regimes more attractive. While the victims in these circumstances could presumably sue the vehicle manufacturer, product-liability lawsuits are more expensive to bring and typically take more time to resolve than run-of-the-mill automobile-crash litigation. No-fault systems are designed to provide compensation to victims relatively quickly, and they do not depend on the identification of an ‘at-fault’ party.”22

But no-fault translates to “no responsibility” and no incentive for technology developers to promote the most rigorous design and testing protocols to ensure their products perform as expected. Even though some manufacturers may accept liability for their AVs, not all will.

It is important to ensure that manufacturers and their products are held up to the scrutiny of the product-specific, strict liability system. Otherwise, the ability to differentiate a safe product from a dangerous one, and the elimination of dangerous products from the market, will be stymied.

A firm and unreasonable demand for preemption. The RAND guide advocates using either implicit or explicit preemption to immunize developers and manufacturers from suit, so long as they adhere to minimum federal standards—standards that they had a hand in formulating.

The guide suggests that Congress preempt state tort litigation, either by flatly limiting liability based on AV technology, precluding legal action where the manufacturer meets regulatory mandates, or by drafting legislation that would implicitly preempt state law.23 RAND claims that the public benefit from AV technology markedly outweighs the risk of harm to some, and cites the 1986 National Childhood Vaccine Injury Act as an example, which limits drug manufacturers’ liability and establishes a no-fault compensation system.24

But not all experts agree. The Brookings Institution rejected these proposed policies and stated that preemption should not be a precondition to the rollout of AVs, noting that products liability law is remarkably adaptive to new technologies.25

The government is not ready to lead, so industry is drafting the playbook. Preemption is an illness, not a cure—it leads to victims subsidizing corporations. Principles underlying the doctrine of strict products liability exist to make corporations accountable and protect powerless victims of defective products.26

Privacy in the AV Society

AVs—as well as modern vehicles’ semi-automated control, information, security, and entertainment systems—threaten our privacy. Onboard computers already record how fast and where we drive, where we stop to shop, the number of passengers, our texts and emails, our contact lists, and the music we listen to.

Manufacturers collect this data in proprietary black boxes, which must be downloaded by a hardwired link to the manufacturer’s own system. Other companies, such as OnStar and TeleAid, wirelessly communicate with vehicles through third-party vendors. Consumers are not consulted, and those who do object and disable the features generally lose other connectivity benefits, such as GPS.

To date, there are no government regulations or industry standards addressing personal privacy in automated vehicle technologies. In one study, researchers found that manufacturers collect, transmit, and sell large amounts of driver data, and customers are unaware of this. The report called on NHTSA to create standards to protect consumer privacy and data. A rulemaking is now underway.27

Exposing the Soft Underbelly

Unfortunately, our quest for convenience has made vehicle hacking much more convenient, too. Control and entertainment systems, including GPS, keyless entry, tire-pressure monitoring, Bluetooth, and Wi-Fi, allow easy access to the vehicle’s central command and control systems.

In July, two cyber-security experts hijacked a late-model Jeep Cherokee from 10 miles away, assuming control of its windshield wipers, radio, ventilation system, and speed—ultimately shutting it down on the highway.28 The hackers gained access through the vehicle’s “UConnect” head unit, which controls the stereo, GPS, and automated driver-assist technologies. The demonstration resulted in Fiat Chrysler Automobiles recalling 1.4 million cars.29

Senators Edward Markey (D-Mass.) and Richard Blumenthal (D-Conn.) have introduced the Security and Privacy in Your Car (SPY Car) Act of 2015, which would require NHTSA, in consultation with the Federal Trade Commission, to develop standards that prevent hacking into vehicle-control systems. The Spy Act calls for final regulations to be issued within three years.30

AV provisions were also included in the Senate’s GOP-proposed six-year Comprehensive Transportation Funding Bill, but the language was removed before passage.31 And in November, Congressman Michael Burgess (R-Texas) introduced an AV amendment to the House transportation bill that would have provided immunity for companies who use “alternative factors or criteria,” to be developed by the manufacturers themselves, to determine whether they have met DOT guidelines. That amendment passed in the House, but was removed from the final bill during the House and Senate conference process.To find out where your state is in the regulatory process, view Stanford Law School’s webpage on AV legislative and regulatory action.32

AV technology has yet to fully manifest, and regulation is just beginning. There is no time to waste.

Christopher B. Dolan is an attorney with the Dolan Law Firm in San Francisco. He can be reached at


  1. Nat’l Highway Traffic Safety Admin., Safety 1n Num3ers (Aug. 2013),
  2. Nat’l Highway Traffic Safety Admin., National Motor Vehicle Crash Causation Survey (July 2008),
  3. Zack Kanter, How Uber’s Autonomous Cars Will Destroy 10 Million Jobs and Reshape the Economy by 2025, (Jan. 23, 2015),
  4. PriceWaterhouseCoopers, Look Mom, No Hands!: Forging Into a Brave New (Driverless) World (Feb. 2013),
  5. Nat’l Highway Traffic Safety Admin., Preliminary Statement of Policy Concerning Automated Vehicles (May 2013).
  6. Greg Miller, Autonomous Cars Will Require Totally New Kind of Map, Wired (Dec. 15, 2014),
  7. Nat’l Highway Traffic Safety Admin., Vehicle to Vehicle Communications: Readiness of V2V Technology for Application (Aug. 2014).
  8. Id.
  9. Id.
  10. Id.
  11. Id.
  12. Vehicle to Vehicle Communications: Readiness of V2V Technology for Application, supra note 7.
  13. Google Founder Defends Accident Records of Self-Driving Cars, L.A. Times, June 3, 2015,
  14. Cal. Vehicle Code §38750 (2015); Nev. Rev. Stat. §482A.030 (2013) (Nevada defines an AV as a motor vehicle, and its definition of a motor vehicle operator can be found at Nev. Rev. Stat. §485.080).
  15. D.C. Code §50-2351 (2013); Fla. Stat. §316.305 (2014).
  16. Nev. Rev. Stat. §482A.090 (2013); Florida Uniform Traffic Control Law, Fla. Stat. §316.86(2) (2014).
  17. Cal. Vehicle Code §38750(b)(3); Nev. Rev. Stat. §482A.060; Fla. Stat. §316.86(1).
  18. Chris Nichols, Liability Could be Roadblock for Driverless Cars, San Diego Union Trib. (Oct. 13, 2013),
  19. Nikolaus Lang et al., Boston Consulting Grp., Revolution vs. Regulation: The Make-or-Break Questions About Autonomous Vehicles (Sept. 2015),
  20. Bryant Walker Smith, Proximity-Driven Liability, 102 Georgetown L. J. 1777, 1818 (2014).
  21. James M. Anderson et al., Rand Corp., Autonomous Vehicle Technology: A Guide for Policymakers (2014),
  22. Id.
  23. Id.
  24. 42 U.S.C. §§300aa-1–34.
  25. John Villasenor, Brookings Inst., Products Liability and Driverless Cars: Issues and Guiding Principles for Legislation (Apr. 24, 2014),
  26. Greenman v. Yuba Power Prods., Inc., 59 Cal. 2d 57, 63 (1963).
  27. Tracking & Hacking: Security & Privacy Gaps Put American Drivers at Risk (Feb. 2015),
  28. Andy Greenberg, Hackers Remotely Kill a Jeep on the Highway—With Me in It, Wired (July 21, 2015),
  29. Andy Greenberg, After Jeep Hack, Chrysler Recalls 1.4M Vehicles for Bug Fix (July 24, 2015),
  30. S. 1806, 114th Cong., §30129(b)(2) (2015).
  31. Surface Transportation Reauthorization and Reform Act of 2015, H.R. 22, 114th Cong. (2015).
  32. The Ctr. for Internet & Soc’y, Automated Driving: Legislative and Regulatory Action,