Self-Driving Cars Are Almost Here, but Questions Remain


How Much Safer Will Self-Driving Cars Be?

Safety is the most oft-cited reason for pushing ahead with self-driving cars. The statistic that safety officials and manufacturers commonly cite is 94 percent—the share of traffic crashes that are the result of human error.

Eliminating human drivers from the equation, they assure us, will lead to a drastic decline in accidents. Yet quantifying just how much safer driving will be without humans isn’t so simple.

Comparing automated vehicle safety to human drivers isn’t as straightforwards as it might seem.

For starters, it depends on which drivers are used for comparison. There are vast differences in crash risk between young and older drivers, alert and distracted drivers, and sober and drunk drivers. It will be relatively easy for an automated vehicles to outperform drivers in high-risk groups. But is a self-driving car safer than someone who has never had an accident in 25 years of driving?

Assuming that a fair comparison between human and machine driving safety can be established, it raises the question of how much safer automated vehicles should be. Twice as safe? Ten times as safe?

Improving traffic safety has renewed importance after 35,092 people were killed in U.S. traffic crashes in 2015—the highest full-year increase since 1966. Experts generally agree that distracted driving resulting from digital technology is to blame for the increase. It is with significant irony, then, that we are looking to technology to save ourselves from the hazards of technology.

In the early stages, the line between human and machine driving safety is likely to remain somewhat blurred as design kinks are worked out. One definite advantage for automated vehicles is the ability to update them. Humans can be told countless times about the dangers of distracted and drunk driving, but never meaningfully adjust their behavior. Machines, on the other hand, can be reprogrammed to avoid making the same mistakes. The more miles driverless cars log, the safer they will become—at least in theory.

Seventy-eight percent of drivers said they’d be “afraid” to ride in a fully autonomous vehicle.

Nevertheless, there is no getting around the fact that imperfect automated vehicles will at some point in the near future be unleashed on roadways. They’ll inevitably be involved in some crashes—along with a bit of “I told you so” rhetoric when they are.

The latter is likely to come from the human drivers that manufacturers are so eager to replace, and who show skepticism about driverless cars.

A recent survey from the American Automobile Association (AAA) found that 78 percent of U.S. drivers said they’d be “afraid” to ride in a fully autonomous vehicle, while 54 percent said they feel “less safe” sharing the road with driverless cars. Only ten percent of drivers said they’d feel safer sharing the road with self-driving vehicles.

No less disconcerting, especially for those who are leery of driverless cars, is whether the public’s opinion even matters to automakers and regulators. And driving enthusiasts might wonder if the day will come when human operation is seen as a dangerous relic that, in the name of public safety, has been outlawed.

The Great Disruption

Safety isn’t the only reason advanced in favor of self-driving cars. In certain circles, the expectations for autonomous vehicles are downright utopian.

Less traffic and pollution, more free time, better health, improved transportation efficacy, and urban renewal are some of the benefits self-driving cars could produce.

However, there are many entrenched interests that stand to lose when human drivers are replaced. The obvious ones include taxi, delivery, truck, and other drivers—who make up nearly three percent of the American work force—as well as traditional auto parts manufacturers, auto repair shops, and gas stations.

Other industries that face big changes include governments (less money from parking tickets and traffic citations), hotels (self-driving cars could be utilized as sleeping pods for overnight hauls), insurance, and the legal profession.

In the legal industry, vehicle collisions account for roughly one-quarter to one-third of all civil trials in a given year. Auto accident personal injury cases are a significant source of revenue for many firms. Because human error is to blame for most accidents, these types of cases will presumably go by the wayside.

Forward-looking law firms recognize that auto accidents won’t be eliminated entirely, particularly in the early going. Cars will crash, and blame will be assigned. Only the blame will be put not on drivers, but on the automotive industry and its suppliers. That is, auto accident law will shift from driver liability to product liability.

“To prove that an automated driving system performed unreasonably, an injured plaintiff would likely need to show either that a human driver would have done better or that another, actual or theoretical, automated driving system would have done better,” says Bryant Walker Smith, a University of South Carolina law professor and technology advisor to the Department of Transportation.

Smith expects that automated vehicles could be held to a higher safety standard, which may result in higher damage recoveries for prevailing plaintiffs. The cost of potential liability, he says, will be factored into an automated car’s sticker price.

Fewer human-caused car crashes also pose a demand problem for the insurance industry. The industry’s solution takes advantage of the shift from human to product liability, proposing coverage that focuses on shielding manufacturers from liability from automated vehicles’ technical failure.

This change, writes McKinsey & Company, “could transform the insurance industry from its current focus on millions of private consumers to one that involves a few [original equipment manufacturers] and infrastructure operators, similar to insurance for cruise lines and shipping companies.”

Flying Cars on the Horizon

Well before regulators and industries have fully adapted to the self-driving car paradigm, they may be faced with another game-changer: flying cars.

Flying cars amplify many of the unsolved problems posed by unmanned aircraft.

More than a dozen large and small companies are working on flying aircraft that could be used for personal and commercial transportation. The involvement of big names like Google founder Larry Page (who is backing Silicon Valley startup Kitty Hawk), Uber, and Airbus lend flying cars legitimacy, but legal and regulatory issues may prove to be more challenging than anything engineering departments face.

Kitty Hawk wants to start selling its flying vehicle by the end of 2017. Uber plans to roll out fleets of driving cars in Dallas-Fort Worth and Dubai by 2020. Airbus plans an initial test flight of its concepts before year’s end.

The Federal Aviation Administration (FAA), which is in charge of air traffic and is still struggling with how to regulate drones, must ensure that flying cars meet safety regulations and fit into air traffic control systems. The NHTSA might also get involved if flying cars use public roadways. Complying with rules from both federal agencies presents a major hindrance for companies—never mind the fact that comprehensive rules for conceptual flying cars are not in place.

Rules will be needed for issues that include: Who will be allowed to operate a flying car? How many passengers are permitted? Can the vehicles fly above roads? What safety features and measures must be implemented?

And that is just for personal flying vehicles.

Under current FAA rules, “ultralight aircraft” cannot be used commercially. That’s fine for recreational aircraft like the Kitty Hawk Flyer but not for the commercial transport flyers that Uber plans. Uber could argue that its vehicle isn’t an aircraft subject to FAA rules. But sorting this and other concerns out in boardrooms and courtrooms could take years.

Private citizens and companies might have separate concerns about their privacy. The FAA has already come under scrutiny for failing to establish rules dealing with possible privacy intrusions from camera-equipped drones. The issue will only deepen when flying cars take to the sky.

Any number of unintended consequences promise to further dampen the flying car dream, to say nothing of the driverless car dream. Yet when one considers that commercial automobiles and commercial air flight are little more than a century old—and that technological change is accelerating—what is possible in the coming years seems almost limitless.

Previous page