A Google prototype of a self-driving car

Congress doesn’t have a reputation for moving fast. On complex topics like health care and taxes, it can take years before lawmakers pass substantive legislation — if they pass any at all.

This fall, however, Congress has moved uncharacteristically quickly to advance legislation governing new technology that is moving quickly: self-driving vehicles.

Automated cars have the potential to upend a delicate regulatory balance: For decades, the federal government and the states have shared power when it comes to the country’s roadways. Washington regulates how vehicles work — like evaluating safety features — while states regulate how the drivers work. But with a self-driving car, the machine and the driver are one and the same.

As the technology behind self-driving vehicles has become more sophisticated — and cars that were once the realm of science fiction began to actually appear on roads, state and local governments stepped in and passed laws to regulate how, and if, self-driving vehicles can operate on their roadways. One effect of this state and local action has been to create a patchwork of different laws, fueling hopes in the automotive and technology worlds that Congress might take action to set federal standards to govern how self-driving cars operate nationwide.

This year, Congress is finally taking on the challenge of figuring out how to navigate a future populated by machines that are both cars and drivers. In September, the House of Representatives unanimously approved legislation to advance a broad regulatory framework for automated vehicles. A companion bill is working its way through the Senate, and it’s possible that compromise legislation could be on President Donald Trump’s desk by the end of the year.

The emergence of a potentially revolutionary new technology presents thorny issues for Congress, where lawmakers want to encourage the growth of the industry while making sure that the public — and their privacy — is protected as robots take over the roads.

A regulatory patchwork

As self-driving cars have gradually rolled out on roadways from California to Michigan, state and local governments have been the first authorities to pass laws regulating them.

Though engineers have been testing prototypes of automated vehicles for decades, a pivotal moment came in 2010, when a fleet of autonomous Toyota Priuses took to the roads around Google’s Silicon Valley campus.

After a big push from Google, in 2011 Nevada became the first state to pass legislation directing its state Department of Motor Vehicles to develop rules that authorized self-driving vehicles and put in place a basic regulatory framework for them.

That law allowed companies to apply for permits to test automated vehicles on Nevada roadways, and it made the state an early focal point for the technology’s development — prompting other states to quickly pass laws of their own.

Currently, 26 states and the District of Columbia have rules in place regarding self-driving vehicles, with different rules and regulations governing operation — and even different definitions of what, exactly, a self-driving car is. (Minnesota does not have any such laws in place.)

There are varying levels of automation: A standard recently adopted by the federal government outlines level zero, where the driver is fully in control of the vehicle, to level five, where there is full automation and no driver is required. Google has said its main focus is on creating a functional level four vehicle, and auto companies like Audi are marketing increasingly automated vehicles that don’t need a driver in some situations, like stop-and-go traffic, that qualify as level three.

But those developments require a lot of testing, and laws to authorize that testing. Proponents of state-level self-driving car laws have tended to frame the issue in economic terms: Governors and state officials have boasted over the years about leading the country with the most innovation-friendly laws.

Michigan Gov. Rick Snyder addressed that dynamic in 2016, when he signed an updated law for his state: “It’s not about racing on legislation,” he said, “but it’s about having good, smart legislation that’s about safety first and then creating an environment for innovation.”

But as more states have developed their own rules, it’s become more complicated for the companies developing the technology — and who want to see it on roads across the U.S. — to comply. For example, California laws require that one person is present inside an automated vehicle as it’s being tested on the road. If that vehicle crossed the border into Nevada, an additional supervisor would need to get inside the car, as that state requires that two people be present during testing.

Congress gets into gear

Since Nevada passed its law in 2011, the federal government has mostly kept states in the driver’s seat on self-driving cars. Federal authorities at the U.S. Department of Transportation have issued annual guidance reports on how they think states should approach the issue, but compliance is voluntary.

But that changed this year, with Congress taking unprecedented steps to set nationwide rules regarding automated vehicles’ safety and operation. In September, the House passed its SELF-DRIVE Act by a unanimous voice vote. The Senate’s AV START Act advanced out of that chamber’s Commerce Committee, and awaits a vote on the floor.

The two chambers’ bills have differences, but both aim to place some central issues surrounding automated vehicles under the purview of the National Highway Transportation Safety Administration, or NHTSA, superseding existing state laws regarding safety and privacy, among other things.

That move would bring automated vehicles into the same regulatory sphere that normal, human-driven vehicles are now: The federal government is tasked with the safety of the machine, and states can regulate most everything else — licensing, registration, and the rules of the road.

Frank Douma

According to Frank Douma, a professor at the University of Minnesota who studies automated vehicle issues, “the federal government always regulates the hardware, and the state regulates the driver, the human. That becomes tricky when you’re looking at the car increasingly becoming the driver.”

In the view of policymakers, the most essential part of regulating automotive hardware is determining how safe a vehicle is to drive, whether that driver be human or not. To that end, the House bill gives NHTSA two years to craft a set of safety standards for self-driving automakers to comply with, and a year to identify exactly which parts of a self-driving car, like laser sensors, will be singled out for rigorous testing.

The bills also lay out a route for more and more self-driving cars to hit the roads in coming years. The House and Senate both allow NHTSA to issue “exemptions” from current federal safety standards, so developers of automated technology can test their vehicles in the U.S. The law allows for 15,000 exemptions in its first year, 40,000 by its second, and 80,000 by its third year. That’s a big increase from the 2,500 yearly permits that NHTSA issues now.

Congress’ legislation recognizes the new challenges posed by connected automated vehicles, and one of the most important ones is privacy. The technology accompanying the cars could store troves of valuable data about their users — detailing where they go and when, for example. The cameras that fully-functional automated vehicles are expected to use could store countless hours of data, too.

The House and Senate bills require companies to explain how they will handle this information, and how they plan to educate customers about their privacy and the use of their data. There are question marks, however, about which government agencies would be tasked with developing and enforcing any rules related to self-driving cars and users’ privacy.

In September, Transportation Secretary Elaine Chao issued a new set of voluntary guidelines for states regarding autonomous vehicles, covering a variety of issues. While guidance from Barack Obama’s administration identified privacy as something states should focus intensely on, Trump’s DOT said that issuing guidance on privacy was not its job — rather, it is the job of the Federal Trade Commission.

Chao’s relegation of privacy issues to a footnote angered consumer advocates, who view self-driving cars as a crucial next frontier on privacy. Lawmakers from both parties in Congress have also said that any regulatory framework needs to clearly address how users’ data might be protected.

Industry allies, meanwhile, have said it’s imperative that the technology be developed as soon as possible — and developed in the U.S., not abroad — and that privacy questions should be solved later.

Supportive lawmakers acknowledge that the legislation isn’t meant to address every question and regulatory hole around automated driving technology; instead, it puts in place a framework that is meant to develop as the industry does.

Sen. Amy Klobuchar, a member of the Senate Commerce Committee that wrote and approved that chamber’s bill, said at a recent Capitol Hill event on self-driving cars that senators intended to strike a balance.

“We’re going to have to do it in a way,” she said, “which is always a tricky thing in Washington, to put some rules in place that allow the innovation to go forward, but don’t try to cross every T and dot every I, because the technology changes.”

Safety first

If the legislation moves forward and gets the president’s signature, much of the nitty-gritty work will lie ahead in the next two years, with government officials left to balance the concerns of public safety and privacy with the pace of innovation.

The industry has given indications of support for Congress’ effort. A team from Maplewood-based 3M, which is working on automated vehicle technology — such as putting technology in road signs to help those vehicles navigate — recently came to Capitol Hill for a panel discussion on the topic.

John Riccardi, vice president of traffic safety and security at 3M, said, “We’re agreed that the lawmakers are taking the right approach here relative to developing policy with respect to automated vehicles.”

Riccardi praised the bills for not being too prescriptive or overly detailed. “The technology is going to change,” he said. “You’ve got to retain that flexibility to allow the technology to move forward.”

Klobuchar and other politicians have raised concerns that a patchwork of state laws — or overly restrictive federal laws — could stifle innovation in the U.S. and could send jobs to countries like China, which could implement more lax rules for self-driving cars.

“If it becomes apparent it’s not going to be easy to deploy these vehicles nationwide in the U.S., folks who are interested in developing them may change the focus to Germany or Japan, even China — somewhere they can have a playing field,” the U of M’s Douma said.

But he added said Congress’ push is a practical opportunity to get some basic safety-oriented regulations done as much as it is a competitive opportunity. “If you saw this kind of unity in any sort of issue, you’d be wondering, what sort of wonderful thing is this?” he asked. “It has to be the safety piece that’s causing everyone to think.”

Safety is the focus back in Minnesota, says Jay Hietpas, the state traffic director at the Minnesota Department of Transportation and the director of MnDOT’s “autonomous bus” pilot project.

Hietpas says it’s a good thing that the federal government is stepping in now to assume its responsibility to gauge the safety of autonomous vehicle hardware, as individual states like Minnesota don’t have the expertise or resources to do that alone.

He adds that now, Minnesota officials are having discussions about how they can fulfill the state-level responsibility to regulate some aspects of automated vehicles as Congress moves forward. That means figuring out, in the short term, what a testing environment might look like in Minnesota — deciding, for example, if automated vehicles need to be marked as such on the roadways.

In the long-term, Hietpas identifies some tough issues — licensing, liability, rules of the road —  that state-level officials are going to have to work through on their own as self-driving cars become more prevalent.

“What are the long-term laws we need to change?” he asks. “Vehicle insurance, licensing and registration, who’s responsible if there’s a crash, who can get a driver’s license — can a 10 year old operate an automated vehicle many years down the road?”

MnDOT will stay tuned, Hietpas says, and does that with some optimism. “We think this is going to help us in Minnesota reduce fatalities and serious injuries in the long term.”

Join the Conversation

3 Comments

  1. This is going to be an interesting transition

    We also may need to re-engineer our roads to help out autonomous vehicles. Example: signage with 3D barcodes, wireless beacons for our cars and trucks (including emergency vehicles) and traffic signals, and new kinds of markers for our road lanes.

  2. it hasn’t even begun

    I’m a former software engineer and I’d say people in government people haven’t begun to imagine how big the regulatory issues will be. Testing one autonomous vehicle by itself only goes so far; they’ll be interacting constantly on the roads, and every car maker will have its own proprietary software.

    If I were in one of their meetings, here are some sample scenarios I’d toss out. I could come up with lots more.

    1. A pedestrian is killed by a Ford, and it turns out the software has a subtle flaw that could be triggered in many common situations. Fords says it will take them 2 months to develop a fix, test it and roll it out to vehicles. Is that good enough? The alternative is to tell the owners of those cars they can’t drive them.

    2. Volkswagen ads start claiming their new model will get you to work 15% faster – by “optimized lane changes”, and “better performance in heavy traffic”. In other words, VWs start edging out other cars, making people angry. Is that legal?

    3. Cities find out that 80% of downtown parking spots are being grabbed by Toyotas, because they’re running a better algorithm on availability data. Word gets out and owners of other vehicles start complaining.

    4. Minneapolis is hit by a surprise winter storm that coats the roads in ice. All the autonomous Chevrolets suddenly start reaching their functional limits, at which point they stop and tell the passengers to call for help. Inside of a half hour, the city has hundreds of people stuck on the highways – with Fords, Toyotas and VWs still trying to maneuver around them. Now what?

  3. Simple Solution

    Ban them altogether. It’s the only option. People must be responsible for their driving, fully responsible and accountable. We cannot have a mix of automated and driven cars. Maybe 100 years from now, when all the ramifications can finally be accounted for, then we could transition, but more likely, we should never do it. It’s just like robotizing all jobs. Even if a car is self-driven, the drivers and riders should still be paying attention, this will only encourage them to be completely distracted, which means more accidents, not fewer.
    We have to start protecting Humanity against Technology, starting now. The Technology fanatics have reached the point of being a danger to society. Already, everything they create is amoral, absent of ethics, inhuman and inhumane.
    If we switched over to paying everyone a stipend just for being alive, what motivation will they have to accomplish anything? It is as debilitating as it is liberating. The only people who need stipends are artists and those who are ill (same thing, in some ways) or worn out.
    The biggest problem is, no one seems willing to think things through and look at the consequences, to make final conclusions, to account for the effects. And manufacturers must be held responsible for the effects of their products, which includes Hollywood. No, everyone wants to dismiss real thinking and just ride the wave, even if it means oblivion.

Leave a comment