So far, manufacturers of self-driving cars have operated in uncharted waters, with regulations governing the rules of the road varying from state to state. Now the federal government is stepping in with a major new effort to bring order to the chaos.
On Tuesday, the federal government released its first rulebook governing the manufacture and sale of self-driving cars — everything from nearly autonomous Teslas to Google cars without steering wheels or foot pedals. Under the new rules, companies that are building and testing self-driving cars will be required to share extensive amounts of data with federal regulators. But it’s unclear that companies will do so without a fight.
Tech companies and automobile manufacturers will have to address a 15-point "safety assessment" — including details on how a car’s software will address ethical and conflict situations on the road — to determine whether or not their driverless vehicles are safe for public use.
The safety assessment is just one aspect of the Obama administration’s hotly anticipated new rulebook for self-driving cars. Last January, US Transportation Secretary Anthony Foxx promised to release a set of guidelines designed to bring some order to the flurry of activity surrounding automated vehicles. Since then, most of those companies have expressed plans to put self-driving cars on the roads — a clear sign that the Obama administration was racing against time with the release of these new rules.
It’s also a signal that the automobiles as we know them are about to change in a radical way. Until recently, all the rules that applied to driving a car assumed that the vehicle’s operator was a human. Now the rules need to reflect a reality in which a computer is controlling the car’s major functions.
"We’re envisioning a future where you can take your hands off the wheel and the wheel out of the car, and where your commute becomes productive and restful, rather than frustrating and exhausting," said Jeff Zients, director of the White House’s National Economic Council, in a call with reporters Monday evening.
The big question is whether manufacturers will willingly share information about their technology and safety record with the feds, or whether regulators will need to be more aggressive in getting companies to disclose this information. A spokesperson for the National Highway Traffic Safety Administration said the guidelines will likely be made mandatory through regulations.
The government is seeking basic details on how these cars function, as well as how they record data, what happens when they crash, how they protect themselves against malicious hacking, and most intriguingly, "how vehicles are programmed to address conflict dilemmas on the road." They plan on publishing the responses they receive in an annual report.
Foxx said the safety assessment was developed in a way to boost transparency around self-driving cars, while also protecting privacy and competitive interests. "It’s in their vested interest to be as upfront and clear and transparent as possible because there’s market risk to putting a product out there that doesn’t meet the expectations of the public," Foxx said of the manufacturers.
That’s not to say the feds will have an easy time getting manufacturers to cough up their data. Earlier this year, executives from Google, General Motors, Delphi, and Lyft were hesitant support legislation to require anti-hacking safeguards in connected and autonomous cars during a Senate hearing. Their squirming under questioning hinted at future conflicts down the road.
Paul Lewis, vice president of policy at the Eno Center for Transportation, said its unclear how rigorous this data sharing with the government will be. "None of these [safety points] have standards associated with them," Lewis said. "It could be that a tech firm will write, ‘We will consider safety.’ Is that enough? It’s hard to judge what the value of this is."
So what else is in this new rulebook? The Department of Transportation and the National Highway Traffic Safety Administration have included a set of model regulations for states to adopt. This is to avoid a "patchwork" of regulations where some states allow for the deployment of self-driving cars, while others do not. This policy governs how to inspect and license self-driving cars, as well as law enforcement and insurance considerations.
Foxx said that government can already exempt manufacturers from including certain features that have typically defined a vehicle, such as steering wheels, pedals, and front-facing seats. But it can also recall any vehicle that doesn’t meet federal safety standards, Foxx said. "We will not hesitate to use our recall authority," he added.
There’s also a list of new powers the feds may seek to acquire to better regulate the coming wave of self-driving cars. This includes a few that are sure to ruffle the feathers of manufacturers, such as pre-market approval of driverless cars and post-sale regulation of software updates. (Earlier, The Verge reported on the government’s blind spot when it comes to modern vehicle software.)
Most of these new authorities would require approval from Congress, as well as more money to hire experts to assess autonomous technology, Foxx said — none of which is a sure bet. By this point, most experts in the field have been snatched up by Google, Uber, Ford, and all the rest. "This has been an area that’s been vexing me for quite some time," Foxx said to a question about his department’s ability to retain the engineering expertise for the work ahead.
Foxx also stressed that the guidelines were meant to serve as an evolving document. There will be 60 days for public comment, and the department will update it annually. Foxx didn’t address any of the public demonstrations of self-driving technology that have occurred so far, such as Uber’s limited trial in Pittsburgh or Google’s fleet of driverless cars in California, Texas, and Washington.
"Our intention is to cover the waterfront as best we can," Foxx said. "There’s going to be a need for conversation over the long term."