CHAPTER EIGHTEEN AUTONOMOUS VEHICLES: DEVELOPMENT PROGRESS AND LACKING LIABILITY SCHEMES

• Jurisdiction: Maryland

CHAPTER EIGHTEEN AUTONOMOUS VEHICLES: DEVELOPMENT PROGRESS AND LACKING LIABILITY SCHEMES

Rebecca K. Schisler-Adams DeCaro, Doran, Siciliano, Gallagher, & DeBlasis, LLP

ABOUT THE AUTHOR

REBECCA K. SCHISLER-ADAMS graduated from Mount St. Mary's University magna cum laude and the Catholic University of America Columbus School of Law cum laude. She is now an associate with DeCaro, Doran, Siciliano, Gallagher, & DeBlasis, LLP. Her practice includes premises liability defense, insurance defense, and criminal defense.

I. INTRODUCTION

This chapter reviews the various stages of development of autonomous vehicles in the United States, with a focus on Maryland and the Mid-Atlantic region. Though still in the early stages of legislative advancement, this chapter presents liability and insurance schemes where applicable.

II. AUTONOMOUS VEHICLE DEVELOPMENT STAGES

A. Development in the Industry

The Society of Automotive Engineers (SAE) has designed a zero to five rating system detailing the varying levels of automation—the higher the level, the more automated the vehicle is—with certain levels still a number of years away from reaching the market.¹ The SAE has determined the following levels:

• Level Zero: No Automation —The driver is responsible to do all the driving without any help from the vehicle
• Level One: Driver Assistance —The vehicle helps steer or speed up/slow down, but the motorist performs all other duties
• Level Two: Partial Automation —The vehicle helps with one or more systems while the motorist does the rest
• Level Three: Conditional Automation —The vehicle completes all duties, but the motorist intervenes when necessary
• Level Four: High Automation —The vehicle completes all driving duties even if the driver does not intervene
• Level Five: Full Automation —The vehicle completes all duties without a driver on all roads in all conditions²

Leading the autonomous vehicle industry are Google's Waymo and Musk's Tesla, but other automobile firms including Ford, Cadillac, Volvo, Lexus, and Navya are developing their own advanced driver-assistance systems.³ As of this writing, even Tesla's "Autopilot," which can aid with changing speeds and lanes, is only at a Level 2.⁴ Other than Waymo, which says it has a Level 4 system, cars with Level 4 and Level 5 features might not be available until 2040 to 2050.⁵

Until a fully autonomous vehicle hits the roadways, features such as cruise control, backup cameras, lane sensors, blind spot alerts, and facial recognition technology give insight into the types of autonomy that may soon become industry-standard.

B. Liability Concerns and Automobile Insurance

1. Who Is Liable?

Traditionally, traffic regulations and automobile insurance assume human control of the vehicle. However, an advancing technological framework creates potential liability for more than just a driver; determining fault may now include considering the owner, operator, passenger, manufacturer, or coder where a human may not be physically present in the vehicle.⁶

In 2018, the U.S. Department of Transportation reported that compliance with federal safety standards "does not automatically exempt any person from liability at common law, including tort liability for harm caused by negligent conduct."⁷ Most states already have tort liability laws in effect holding both the driver of the vehicle responsible for any accidents they cause as well as holding manufacturers responsible for any defective product. Where there is tension between human causation and product reliability, tort liability law lacks support.

In general, "most litigation surrounding car accidents revolves around a driver's alleged negligence," and requires proof of a duty, breach, causation, and harm to another.⁸ Notably, negligence requires neither proof of intent nor reckless disregard; negligence merely requires proof that conduct fell below a reasonable standard.⁹ The question, then, becomes whose actions do we deem unreasonable? Autonomous vehicles are paving the way for tort reform in the area of negligence and respondeat superior liability. Theories of negligent hiring and supervision could be utilized to hold a manufacturer liable for accidents where a human was present, even when the accident was not caused by any present vehicle defect.¹⁰

Within the context of tort reform, an emerging school of thought presents a concept called "AI-Chaperone Liability."¹¹ This occurs when:

(1) One who acts as a chaperone for AI that is performing a traditionally human task 'is subject to liability for physical harm thereby caused' to another or another's property, if

(a) the chaperone has either supervisory authority over or may override the AI device/system, and
(b) the one harmed is responsible for any part of the harm experienced, then damages may be reduced.

(2) 'The rule stated in Subsection (1)' uses the following definitions:

(a) 'artificial intelligence' is 'the capability of a machine to imitate intelligent human behavior';
(b) 'chaperone" is any entity 'who accompanies and looks after' the AI device/system; and
(c) 'traditionally human task' is any behavior that was performed by a human before the automation of the function.¹²

The purpose of this emerging concept is to fill the gaps between intentional torts, negligence, and products liability that artificial intelligence presents.¹³ While keeping technology at the forefront of the conversation, AI-Chaperone liability embraces the idea that a human is still the mechanism by which autonomous vehicles must operate.

It does not appear that states have yet embraced tort reform in the manner that AI-Chaperone liability encourages, as most states are still in the testing stage of introducing autonomous vehicles. But as states begin to navigate the fine line between holding manufacturers accountable and taking into account human intervention, AI-Chaperone liability could be a solution.

Where states have chosen to advance autonomous vehicle liability legislation, many have chosen to protect the manufacturer. Some states, including Michigan and Nevada, "have limited manufacturer liability by stating that manufacturers cannot be held liable for any unauthorized third-party modifications to their vehicles."¹⁴ Similarly, some states have implemented new definitions of the word "operator" to protect manufacturers; Texas, for example, holds the owner of the vehicle responsible for accidents and traffic violations.¹⁵ Tennessee, on the other hand, requires that the manufacturer assume liability for any instances where their automated driving system is in control.¹⁶ Arguably, the choice to protect the manufacturer is calculated and purposeful—a state that places less liability on a corporation opens itself up to becoming a hub for innovation.

2. Changes in the Insurance Landscape

There is no doubt that these changes in the law will spark change in the insurance industry as a whole. Manufacturers of autonomous vehicles promise that the cars will be "increasingly safer, which will significantly reduce accident frequency and, potentially, accident severity."¹⁷ In 2017, KPMG estimated that the changes in safety brought about by autonomous vehicles could save the industry from $122 billion in aggregate losses.¹⁸

Experts predict three key changes to the insurance industry in upcoming years:

First, we can expect a shift to manufacturers of driving risk and associated liability as more driving decisions are made by a vehicle's proprietary algorithmic "brain." Volvo and other manufacturers have announced that they would accept responsibility and liability should an accident occur due to their autonomous technology.
Second, in an environment where driving decisions are shared between the driver and the vehicle, we may see a consolidation of legal exposure. Providing insurance to both the driver and the manufacturer could prove to be a legal advantage by reducing the volume of cross-suits between the two regarding blame.
Third, the next generation of cars will capture increasing amounts of data via an array of sensors and cameras. A key issue to be determined by legislation and court rulings will be identifying which parties will have the legal authority to use and control that data.¹⁹

Given...