Here’s how it usually happens: Technology comes first, and the law follows. Airplanes first flew across the skies long before control towers popped up, and cellphones were ubiquitous long before texting-while-driving was outlawed. This makes sense. It’s difficult, bordering on impossible, to predict all the ways in which transformative new machines will change the world.
That doesn’t, and shouldn’t, stop people from trying. U.S. Department of Transportation officials on Tuesday unveiled long-awaited policy guidelines for self-driving cars, issuing a sweeping document that reveals how the nation’s top transportation leaders plan to regulate new technology that promises to make the roads vastly safer.
The guidelines clearly demonstrate the federal government’s support for driverless cars, and serve as a regulatory blueprint to “accelerate the HAV [highly autonomous vehicle] revolution” safely and consistently across the nation, the document says. (The agency is also working with officials in Canada in Mexico to suggest similar regulatory framework, it says.) Although most of the new policy is effective upon publication, the Department of Transportation is still seeking public comment in response to its guidelines, and says it will hold a series of public workshops to gather additional public input.
Manufacturers, so far, have responded to the guidelines with cautious optimism—with many of them highlighting the onus on states to consider and implement complementary policies. “This will be a crucial journey that can only be accomplished by working together to avoid a patchwork of policies that could stymie technological development expected to someday save tens of thousands of lives per year,” said Scott Keogh, the president of Audi of America, in a statement provided by a spokesperson.
“State and local governments … should work with the federal government to achieve and maintain our status as world leaders in innovation,” said David Strickland, the head of the Self-Driving Coalition for Safer Streets, which represents a cohort of companies like Google and Uber.
Self-driving cars have become “the archetype of our future transportation,” officials wrote in the National Highway Traffic Safety Administration (NHTSA) report. “Still, important concerns emerge. Will they fully replace the human driver? What ethical judgments will they be called upon to make? What socioeconomic impacts flow from such a dramatic change? Will they disrupt the nature of privacy and security?”
In addition to tackling many of the complex, unprecedented questions that self-driving cars raise—about user privacy and ethics, for instance—the federal government is radically reconsidering its own authority as it pertains to the regulation of automobiles.
In the era of driverless cars, the Department of Transportation is poised to take on a bevy of new responsibilities such as inspecting new vehicles, testing and regulating self-driving software, and creating new kinds of vehicle-testing environments—essentially establishing a new kind of federal DMV focused on driverless vehicles. (Under the new guidelines, states would still be responsible for issuing driver’s licenses to humans, handling registration and titling, carrying out enforcement of traffic laws, and regulating insurance and liability.)
The question of how to create effective policies for driverless cars—including licensing them—has been at the center of debate about the new technology.
“We have to figure out how to do a certification test,” Missy Cummings, the head of Duke University Robotics, told me last spring. “How do we certify humans? Well, we take them all in, we give them a stupid test—written and driving. The idea [for driverless cars] is that we take a car, and run it through a set of scenarios. If that same software is on the other million cars, then great. That car is ready to go, right? So we just need a new DMV for driverless cars. Because you’re not going to be able to test the code line by line.”
All this is happening at a time when state-level DMVs are themselves transforming. “When we talk about what’s coming in the future with autonomous vehicles, that’s only a continuation of the real evolution that we saw really start with 9/11,” said Jenny Openshaw, a vice president at MorphoTrust, an identity services company. “Before that, DMVs were really only focused on making sure drivers had the requisite skills to drive a car. After 9/11, there was a sudden realization that the processes and procedures [for someone] to prove that they are who say they are, were more important than ever. Really, what we’ve seen, is the DMV being more than just a driver’s license agency, but also an identity bureau.”
“A lot of the commissioners of DMVs across the country would tell you their staff spend equal if not more time on identity profiling as they do on testing drivers for knowledge,” she added. “We think that’s its very possible that the identity mission for the [state-level] DMV will actually become stronger than the certifying mission.”
It’s conceivable, in other words, that the licensing of human drivers as we know it will very soon become outdated. “As more self-driving cars are tested and deployed, local departments of motor vehicles will no longer need to license human drivers,” wrote Johana Bhuiyan for the technology news site Recode. “The software will be driving the cars, so long as a 15-point safety standard has been met.”
Kurt Myers, the deputy secretary for Driver and Vehicle Services in Pennsylvania and one of the leaders of a task force on driverless-car policy recently put it to me this way: “There’s never been anything quite like this, other than the introduction of the automobile.”
Back then, it took years before licensing was a requirement. This lag was despite carnage on the roadways, where automobiles had become a “serious menace to life and limb,” as The New York Times reported in 1899. (The public safety threat remains: 35,200 people died in traffic accidents last year, a startling 7.7 percent increase compared with the year before.)
In the 19th century, the best way to regulate automobiles was a contentious issue—similar to the debate around driverless cars today. Beginning in 1901, in New York, getting a license simply required mailing a letter that included your name, address, and a description of your car with a $1 fee to the Secretary of State. That licensing law also came with a collection of other new rules:
All automobiles had to have a bell, horn, or “other signal.”
All automobiles had to have two white headlines and a red tail light.
All automobiles had to have a license plate bearing the initials of the owner—displayed at least three inches high.
The new speed limit was 8 miles per hour.
Oh, and then there was this rule, too: Every person driving an automobile was required to “immediately stop and remain stationary” when encountering someone “driving or riding a restive horse or horses or driving other animals.”
In the 115 years since then, things have changed dramatically on America’s roadways—and they’re about to change again.
“New vehicle technologies developed in the 20th century—from seat belts to air bags to child seats—were once controversial,” officials wrote in the NHTSA report. “But after having saved hundreds of thousands of American lives, they are now considered indispensable.”