Chapter 63 -- Liability

“The truth? I thought we were talking about a court of law. Come on, you’ve been around long enough to know that a courtroom isn’t a place to look for the truth.”
From “A Civil Action”85

The history of the automobile is also the history of a steady socialization of risk and liability. Early drivers risked their lives and property when they took to the road and eventually a steady stream of accidents forced society to demand some form of accountability. States started requiring insurance or some sort of bond that would protect the victims of any mistakes on the road.

Spreading risks and ensuring the safety of riders and pedestrians is the biggest challenge facing the emerging autonomous industry. Self-driving cars have already driven hundreds of thousands miles without accident but no one is sure how to build a legal environment where they can thrive. On one hand, the software will never get drunk or lose vision with age. When mistakes are made, there will usually be a good video stream to settle the matter without excessive debate and litigation. (See Chapter 64 )

On the other, computers and the people who program them are far from perfect and no one knows just how autonomous cars will fail. They may make small, easily fixable mistakes or create dramatic, system-wide deadlocks that leave the country in total gridlock. No one knows what to expect and that means no one knows how to plan or budget or insure such a potential danger.

Start up companies are often good at playing the legal game. If they choose poorly, they just go out of business. If the software fails dramatically, investors lose everything and victims may or may not get some compensation. Risk taking is what startups do well and there will be a role for startups in this marketplace.

Google is no longer a startup but it brings a unique set of skills to the problem. Their technological prowess is certain, but it may be their legal team that allows their car to thrive. In the past, they have shown a willingness to push forward with technology even when laws are uncertain or forbidding. The company has ignored many of the potential liabilities of widespread copyright infringement and created a business that thrives in spite of what many see as widespread lawbreaking.

Companies that start selling automated cars will need this kind of attitude. Already, Google has lobbied California and Nevada to broaden their definition of driving to include simply giving the computer a destination. If the car makes a mistake interpreting these instructions, then someone can still sue the driver or, most likely, the driver’s insurance company. In other words, it’s still sort of the driver’s fault.

This isn’t as much of a stretch as it seems. When a driver turns the wheel on a power-steering car, the driver is trusting that the car’s mechanism will follow the instructions. The computer is just a fancier, smarter version of power steering. At least that’s how some philosophers would like to imagine it.

There will also be lawsuits against the programmers and the testing companies that make the cars, If the software in the cars is anything like the software in my laptop, the cars will be wonderful until they just stop working for some inexplicable reason. Restarting them will often fix the problem, but this isn’t an option when you’re blasting down the freeway going 95 mph.

The trickiest problem will be figuring out how to assign blame when both human and computer are less than perfect. If the tires are getting worn and lose their grip on a rainy night, the driver is supposed to be smart enough to fix the problem before it happens. But if the tire explodes because it is underinflated and the driver is going too fast, no one knows whom to blame.

The general legal environment for cars has been evolving away from blaming anyone or anything. More and more insurance plans are said to be “no fault” and the companies simply pull the money to fix things from a big pool to which everyone contributes.

Autonomous vehicle fleets will most likely take on all of the responsibility. They will effectively bundle insurance with the ride and the only question will be how much they will pay for mistakes.

The rise of the fleets will also largely fix the small but unnerving problem of uninsured drivers, a problem that defies the ability of the legal system. If people aren’t going to listen to the laws about insurance, they’re not going to listen to other laws. But if you don’t have the money to afford the increasingly expensive insurance, you’ll take your chances. As Janis Joplin sang in “Me and Bobby McGee”, “Freedom is just another word for nothing left to lose.”123

The arrival of a fleet will slowly reduce this problem by offering people the chance to book rides without saving enough money to own their own vehicle. (See also Chapter 6 ) This pay-as-you-go system is common for mobile phone users who don’t have the money to handle a two-year contract.

The fleet will offer a tempting, instant solution that will socialize this liability and make the roads safer.

Another solution is for the government to craft some risk pool. The U.S. government, for instance, handles risk pools for vaccines, floods, and a few more events. If the new marketplace for rides is open to all, it’s easy to make the argument that the government might manage the insurance and share the risk. After a few years, we’ll understand the risks well enough and the government will probably be able to step out of the loop.