R2DFORD: autonomous vehicles and the legal implications of varying liability structures.

• Author: Herd, Alexander

Abstract

The World Health Organization estimates that by 2030 traffic accidents will be the fifth leading cause of death in the world. Thus, when Google announced that it had designed an autonomous car that could reduce traffic accidents by as much as ninety percent, there was cause for excitement. Some states have already enacted legislation to permit the use of autonomous cars in anticipation of their release later this decade. As a result, courts and lawmakers need to consider who will be liable when the car that drives itself crashes.

Standards used in aviation and naval cases regarding autopilot can be applied to the more advanced technology behind autonomous cars using the pilot-in-command regulations. Product liability standards, such as the "consumer expectations" and "risk-utility" tests, may also provide guidance. The complex nature of the autonomous system and the differences between a pilot using auto-pilot and a driver allowing the car to drive itself make the "pilot-in-command" and "consumer expectations" standards less than ideal. This Comment argues that the flexibility and balance from the risk-utility test is the most suitable fit for the autonomous car.

1. INTRODUCTION II. BACK TO THE FUTURE: CURRENT LAWS AFFECTING LIABILITY FOR THE AUTONOMOUS AUTOMOBILE A. The Technology Behind the Autonomous Car and the Economic Impact it Could Have B. Initial Lawmaker Response to the Google Car and Issues of Legality C. Consumer Expectations and Risk-Utility Standards D. Aviation and Naval Guidelines for Auto-Pilot III. HOW THE AVAILABLE STANDARDS CAN BE APPLIED TO THE AUTONOMOUS CAR A. Applying Consumer Expectations: What to Expect When You're Expecting a New Technology B. Risk-Utility Applied: Discussing the Benefits of Applying the Standard C. For the Autonomous Car, Does Aviation Law Fly or Can Naval Standards Anchor the Analysis? IV: COURTS AND LAWMAKERS SHOULD CONSIDER THE PURPOSE OF AN AUTONOMOUS CAR AND THE IMPACT ON SOCIETY WHEN DECIDING WHICH STANDARD TO APPLY A. In a Perfect World: Potential Societal Impacts of a World Filled with Autonomous Cars and How Each Standard of Liability has an Impact B. The Risk-Utility Standard is the Most Suitable for the Intent Behind the Autonomous Car V. CONCLUSION I. INTRODUCTION

   What if there was a car that saved money on gas, kept occupants safer, and gave the driver more free time from commuting? This dream may only be a few years away from becoming a reality. (1) In 2010, Google announced that the era of the autonomous car was quickly approaching. (2) Google believes the introduction of its cars can significantly cut down on the number of accidents, car usage, and energy consumption. (3) The company even predicted that its autonomous cars might be able to cut the number of traffic fatalities worldwide by half, and by some estimates, even up to ninety percent. (4)

   While it will likely be a few years before the autonomous car is sold to consumers, (5) those interested in the technology are beginning to think about who would be liable if a car driven by a computer gets into an accident. (6) One suggestion is to apply current aviation law to the driver of an autonomous vehicle. (7) Others believe that existing product liability standards, such as the "consumer expectations" or the "risk-utility" tests, can be applied to autonomous technology. (8) Aviation standards may make assigning liability significantly easier but could require the car to be used in ways other than how Google intended. (9) However, products liability tests may not be applicable to autonomous vehicles until they have been on the market for a number of years. (10)

   Three states have approved legislation that allows the Google autonomous car on their roads. This legislation will also begin a process of incorporating the car into current vehicle codes. (11) For example, California's law mandates that the State's Department of Motor Vehicles have regulations in place for autonomous cars no later than 2015. (12) Still, more states will need to pass similar legislation for the car to be successful. Even so, the current traffic regulations set forth by the Geneva Convention, the National Highway Safety Administration, and state codes would likely not prohibit the use of an autonomous car. (13) However, states need to make room in their codes for the autonomous car to avoid complicating the process for the courts. (14)

   The choices that lawmakers and courts face in assigning liability could have major ramifications on Google's ultimate decisions regarding what could potentially be a trillion dollar product. (15) For example, Google may need to consider offering training on the proper use of the autonomous system to every new purchaser. (16) In order to prepare for suits that arise out of system malfunctions, Google may also need to collect and be prepared to present data supporting its position that the benefits of the autonomous car outweigh any potential safety hazards. (17)

   This Comment analyzes the end result by applying different liability structures to the use of an autonomous car. Section II first summarizes the consumer expectations and risk-utility tests used in product liability law. The section then discusses the standards for auto-pilot, including the pilot-in-command standard of the Federal Aviation Administration ("FAA"), currently used for planes and boats. Next, in Section III, this Comment examines how a consumer expectations test can be applied to the Google car and the consequences that could result from using the test. This section also describes the protocol for applying the risk-utility test to an autonomous car and explains the steps both a plaintiff and Google would need to take during a lawsuit. Additionally, Section III discusses the pilot-in-command rule and other aviation standards as they are applied to the autonomous vehicle. Finally, in Section IV, this Comment argues that the risk-utility standard results in the best outcome for drivers without releasing manufacturers from all responsibility for their product.

2. BACK TO THE FUTURE: CURRENT LAWS AFFECTING LIABILITY FOR THE AUTONOMOUS AUTOMOBILE

   This section explains the technology used in the Google car and the benefits its design provides to many drivers. Then, it discusses the reaction of state lawmakers to the autonomous car, as well as its legality under current United States law. Finally, this section summarizes and explains the relevant liability standards used in aviation and naval law, including the consumer expectations and risk-utility tests.

   1. The Technology Behind the Autonomous Car and the Economic Impact it Could Have

      Google has recently revealed how its autonomous car is capable of driving without human control. (18) The key mechanism of the car is a laser positioned on the roof of the car which creates a three-dimensional map of the environment. (19) This laser produces an image which is then "combined with ... high-resolution maps programmed into the car." (20) The car is also equipped with four radar systems used for far-sighted vision on the freeways, as well as a camera for monitoring stoplights. (21) Additionally, the car has a Global Positioning System ("GPS"), "an inertial measurement unit, and wheel encoder." (22)

      To aid the car's artificial intelligence in determining whether the object is a pedestrian or some other stationary, inanimate object, Google engineers developed a system that allows the car to compare its current environment to one preloaded by Google. (23) Google engineers drive along a route to gather data about the environment. (24) Then, the car travels the same route and simultaneously compares the new data with the data gathered by Google engineers; this comparison allows the car to determine if a pedestrian is present. (25) The car is also programmed to behave more aggressively in certain conditions, such as inching forward to show intention. (26) However, despite the many technological features, certain conditions, such as ice or heavy rain, may render the car's technology unusable. (27) If these conditions are present, the car is designed to alert the driver that it cannot use the autonomous technology to drive, which compels the driver to take over operation of the vehicle. (28)

      Even though the excitement surrounding the technology has started to build, there is the looming potential for enormous social impacts. (29) Sebastian Thrun, lead developer of the Google car, claims the car could result in reducing traffic accidents by 90%, wasted commute time and energy by 90%, and the number of cars by 90%. (30) In 2009 alone, car crashes contributed to over 33,000 deaths and two million injuries in the United States. (31) The American Automobile Association estimated the total cost of car-related accidents in the 99 largest U.S. urban areas was $299.5 billion; this figure suggests an estimated cost of around $450 billion for the entire country. (32)

      If Google's claims are accurate, in a year's time the autonomous car will save 30,000 lives, prevent two million injuries, and cut collision related expenses by at least $400 billion. (33) Improved safety on the roads is a major concern. The World Health Organization ("WHO") estimates that traffic injuries will become the fifth leading cause of death worldwide by 2030, accounting for 3.6% of total deaths. (34) With a product that could save billions of dollars worldwide, knowing how to best prepare for potential liability is vital to how Google conducts business.

   2. Initial Lawmaker Response to the Google Car and Issues of Legality

      Before consumers can purchase autonomous cars, lawmakers need to determine if operating the car is legal. To answer this question, the Geneva Convention, National Highway Traffic Safety Administration regulations, and state vehicle codes need to be analyzed. (35) The Geneva Convention, which the United States is a party, establishes uniform rules for road safety; one of these rules mandates that...