The Department of Transportation is planning to publish the first-ever federal guidelines on how to regulate the deployment of self-driving cars tomorrow.
The guidelines, which will appear on the DOT’s website, are the government’s first attempt at addressing both safety concerns over self-driving technology, as well as the concerns of companies like Uber and Tesla that are eager to get the fast-moving technology into the hands of consumers.
The National Highway Traffic Safety Administration and DOT crafted the guidelines based in part on public feedback after a series of hearings. There, representatives for the companies that are aggressively investing both time and money into developing self-driving technology emphasized the need for a consistent set of national laws on how to deploy the highly automated vehicles, rather than a plethora of confusing state laws.
"The worst possible scenario for the growth of self-driving vehicles is an inconsistent patchwork of laws," Robert Grant, Lyft’s director of government relations, said during his testimony at a NHTSA public hearing in April. "Regulations are necessary, but regulatory restraint and consistency is equally important if we’re to allow this technology to reach its full potential."
Representatives from DOT and the National Economic Council made it clear on a call with reporters today that they are also trying to avoid inconsistent state regulations, because those could inhibit the national deployment of self-driving technology and potentially give other countries a leg up over the U.S.
“States are welcome to do what they want to do, but we hope it will be in support of a [consistent uniform national policy],” said Jeff Zients, director of the National Economic Council.
While DOT has made its position clear, it’s still up to states to decide many of the specifics of their own regulations. That structure, however, could easily lead to inconsistencies in laws from state to state.
To clarify the road ahead, the agencies provided a fact sheet Tuesday that outlined some of the basic tenets of the guidelines, but didn’t go into too many details.
For example: States could control whether a company can test its vehicles on public roads, how to enforce these rules, the method for registration and titling of self-driving cars and, perhaps most importantly, the issue of liability and insurance.
The remainder of licensing and traffic laws will work the way they do today, but the federal government will take over licensing self-driving software based on whether it meets the safety standards it has set forth. As more self-driving cars are tested and deployed, local departments of motor vehicles will no longer need to license human drivers. The software will be driving the cars, so long as a 15-point safety standard has been met.
The safety standard is not so much a prescriptive approach, where companies would have to meet certain proof points set by the agency, DOT secretary Anthony Foxx said on the call, but an open-ended approach in which each vehicle would be evaluated independently to determine whether it meets a certain objective.
It’s left open-ended, Foxx said, because it takes into account the varied approaches many automakers and tech companies are taking in developing self-driving cars.
There’s little detail included in what the DOT provided to reporters, but among the 15 factors the DOT will evaluate in deciding whether a self-driving car is safe enough are:
- How and where the car is supposed to operate.
- How the car detects and responds to oncoming objects.
- How the system responds in case of a software failure.
- How it was tested and validated.
- How it protects user data.
- How it records and shares data.
- How it is programmed to address ethical dilemmas on the road.
The guidelines also outline a data-sharing mechanism that would allow all automakers and companies to exchange information that would help autonomous cars operate more efficiently.
It’s a strategy that expands on what many self-driving startups and now Tesla see as the future of autonomous technology: Fleet learning. Tesla founder Elon Musk’s premise was that if one of its cars detects an object and is able to drive around it using Autopilot, then the rest of the fleet of Teslas will know how to self-drive around that object. That same logic applies to the rest of the industry.
The goal is a robust platform that is being informed by all self-driving cars and is informing all self-driving cars.
The Self-Driving Coalition for Safer Streets, a lobbying group that includes Ford, Google, Volvo, Uber and Lyft, sees these guidelines as a productive first step but is keen on working closely with states to create consistent regulations.
“State and local governments also have complementary responsibilities and should work with the federal government to achieve and maintain our status as world leaders in innovation,” David Strickland, the coalition’s spokesperson, wrote in a statement. “With the guidance now publicly available, we encourage state policymakers to engage with our Coalition to develop the appropriate policy solutions, and we stand ready to provide support and expertise for both technological and policy questions.”
Many of the policies go into effect as soon as the guidelines are published tomorrow morning. For those stakeholders and players dissatisfied with the new guidelines, the NHTSA and DOT plan to update the policies annually based on public comments.
This article originally appeared on Recode.net.