Historically, the Department of Transportation has been a bit of a backwater. Its secretaries haven’t wielded the power or gained the notoriety of a secretary of state or secretary of the Treasury. But Elaine Chao, Donald Trump’s pick for the department, could prove to be the most consequential secretary in the department’s history. Under her leadership (if confirmed), the Transportation Department will be making some of the most important decisions in the American economy.
The reason is simple: Robots are coming to the transportation sector. Cars and trucks will start to become self-driving. Autonomous drones are going to proliferate. And software-driven innovations like flying cars will make big strides.
Right now federal regulations in all of these areas are woefully inadequate. If Chao is confirmed, she will lead the Trump administration’s efforts to change that.
The big question she’ll need to wrestle with is how to regulate software that has the potential to kill people. Nobody has ever died from a bug in Angry Birds, but flawed software in a self-driving car or drone can be lethal. At the same time, too much regulation could slow the rapid pace of innovation traditionally enjoyed by the software industry.
If Chao is able to thread this needle, she will preside over a historic transformation in the way Americans move themselves and their stuff around. But if she screws up, she could hamper progress and cede the technological edge to foreign competitors.
The next four years will be crucial for self-driving cars
2016 was the year major automakers got serious about the threat — and opportunity — posed by self-driving technology. Uber and Google already have self-driving cars on the roads, and most car companies have announced plans to test their own vehicles on public roads before the end of 2017. Ford and BMW have even pledged to introduce fully self-driving cars by 2021.
The flood of self-driving car prototypes onto America’s streets will give regulators at the National Highway Traffic Safety Administration (NHTSA) a sense of urgency about the need to update applicable regulations. NHTSA is part of the Department of Transportation, and It will need to decide whether to establish uniform national safety rules for self-driving cars — or whether to allow states to set their own divergent standards.
NHTSA published guidelines last September setting out the Obama administration’s vision for the future of self-driving cars. But these were only guidelines; they aren’t legally binding regulations. As a result, many of the rules that will govern self-driving cars are being written by the states.
The optimistic view is that state-by-state regulation will allow states to experiment and learn from each other, but it could also lead to a crazy quilt of rules that make it very difficult to build a car that’s legal in all 50 states.
If NHTSA does decide to step in, it will have to figure out the best way to craft and enforce safety rules for computer programs. It’s fairly easy to measure if a car’s brakes or airbags meet minimum technical standards. But it’s much more challenging to determine whether a computer program is free of life-threatening defects, because software that seems to work perfectly in ordinary circumstances might suddenly malfunction when it encounters an unexpected situation — or when targeted by a hacker.
But Eli Dourado, a researcher at the Mercatus Center, worries that safety concerns could lead to excessive regulation of self-driving technology, slowing its introduction to the market. Conventional cars kill about 30,000 Americans per year, and many of these deaths are caused by drivers who are intoxicated, distracted, or driving on too little sleep. So substituting these drivers for a computer program that can drive as well as the average alert driver would save thousands of lives.
In particular, Dourado warns against requiring companies to submit software updates for prerelease approval by federal regulators — a requirement that could slow progress to a crawl. Instead, he argues, regulators should establish general technical requirements and then allow companies to do their own testing and certify that their vehicles meet the requirements.
Chao could shape the future of drone deliveries
Last year, the Federal Aviation Administration released new rules for unmanned aerial vehicles — popularly known as drones. The new rules lifted a previously strict ban on using drones for commercial purposes. But they also left in place rules that could significantly stunt the development of drone technology over the next decade.
One of the most important rules requires that drones stay within view of a human pilot. It might seem like common sense to have a human being monitoring a drone as it flies around. However, the requirement rules out what could be the most significant drone application: automated deliveries.
Amazon first unveiled its plans for a drone-based delivery network in 2013, and it has been working on the plan ever since. In 2015, the company published a pair of white papers laying out its vision for how drone delivery networks should be managed in the future.
Amazon envisions a future where thousands of small vehicles are buzzing over our heads at any particular point in time. One of the big challenges with that vision is how to prevent them from running into each other. Conventional airplanes do it with a human air traffic controller, but that obviously won’t work in a future where there are many more vehicles in the sky.
Instead, Amazon wants the FAA to require drones that fly above 200 feet to have automated collision avoidance and communication technology on board. Autonomous drones would need a GPS and highly accurate geospatial data, a reliable internet connection, the ability to communicate wirelessly with other drones, and sensors to detect and avoid collisions with other objects in the sky.
The Obama FAA hasn’t endorsed these ideas yet. Indeed, it only opened up the sky to commercial drones at all after Congress insisted that it do so. Which means this is an area where Secretary Chao could make a mark.
But as with self-driving cars, crafting the right regulations won’t be easy. If the FAA proceeds too cautiously, it could stunt the development of the drone delivery industry. If Chao moves too fast, it could lead to drones falling out of the sky, causing property damage and even deaths. Moving too fast could also freeze bad standards into place, preventing subsequent progress.
Chao could nurture the development of flying cars
The technology to create unmanned drone delivery services exists today; the technology is largely being held up by regulation. But eventually, similar technology will likely open the door to flying cars: small, automated electric airplanes capable of taking passengers on short trips within the same metropolitan area. These planes can take off and land vertically, eliminating the need for airports with long runways. And if they can fly themselves, they’ll be open to anyone, not just people with pilots’ licenses.
As I wrote a couple of weeks ago, this technology is still in the experimental stage. One of the big obstacles is that today’s batteries don’t quite have enough energy density to allow electric airplanes to fly long enough distances to make the technology truly useful. There’s also room for further progress on aircraft design and control software.
But there are now several well-funded startups working to overcome these obstacles. Uber has also taken an interest in this market, writing a comprehensive paper arguing that you’ll eventually be able to hail a tiny electric airplane with your smartphone the same way you can hail a car today.
But federal regulators are nowhere close to having a legal framework for regulating this kind of self-flying airplane. Allowing tiny drones to fly without a human pilot is one thing, but allowing airplanes controlled by software to carry human passengers is a whole different level of risk. It’s a level of risk that the risk-averse FAA probably won’t want to stomach.
Still, there are steps the FAA could take steps to promote this emerging market. Commercial flights are probably out of the question in the short run, but the FAA could make sure there’s a legal framework in place for companies to test out their prototypes. And the FAA could begin studying issues like how it might manage the airspace over a city in a future where flying cars are commonplace.