Uber’s Secretive Self-Driving Trucks Could Undermine the Industry

We can’t trust self-driving car efforts until we have greater transparency.

One of Uber’s self-driving trucks.
One of Uber’s self-driving trucks.

A number of lawmakers in our nation’s capital have been wary of self-driving cars on our roadways, a sentiment echoed by many citizens dubious of the burgeoning technology. But while national-level regulation has been slow to come, autonomous car development is already quite advanced in states like Arizona, where self-driving vehicles may even be further along than previously thought. Thanks to a lax 2015 edict by Arizona’s governor, the state has become a hotbed for Silicon Valley’s self-driving car testing. Waymo, one of the leaders in the autonomous car space, recently put in the paperwork to launch a self-driving ride-hailing business in the state. Its biggest competitor has made progress of its own: On Tuesday, we learned Uber already has autonomous trucks driving on Arizona’s highways.

For several months, Uber’s freight trucks have been making runs across Arizona, thus far with drivers still sitting behind the wheel in case manual intervention is needed. In the future, that won’t be the necessary. In a report by the New York Times, Uber explains that it imagines its self-driving technology will handle long-distance hauls, while human drivers will take the wheel for short pieces of the routes, such as navigating the last few miles from a freeway into a loading dock.

Uber has been steadily stepping up its autonomous car efforts, despite a problematic intellectual-property battle with Waymo that it settled in court last month. Uber’s endeavor largely began in 2016, when the company acquired self-driving truck startup Otto. In the first public test of its technology, Uber sent one of the trucks on a 120-mile trip through Colorado to deliver beer. Since then, we’ve learned of Uber’s efforts testing vehicles in Pittsburgh, Tempe, Arizona, and San Francisco, and about its plans to expand its fleet in the coming years. But unlike Waymo, which issued a self-driving safety report on its efforts and gives regular insights into its progress, Uber’s updates in the space have been less transparent—and that’s troubling.

With Uber’s foray into autonomous freight deliveries, for example, the New York Times notes that there are a lot of unknowns. Uber hasn’t revealed how many trucks are currently being tested out on roadways, the number of miles they’ve driven, or the number of times human truck drivers have had to take the wheel during these runs. For an experiment running on public roadways—that could feasibly pose a danger to other drivers if something went wrong—the public deserves more information.

As a company, Uber has a history of ethically questionable moves, including spying on competitors, tracking the locations of competitors’ drivers, and using a tool to remotely conceal data from investigators on more than two dozen occasions—a revelation that led to a federal criminal investigation. While its autonomous vehicles haven’t been at fault in any serious accidents, they don’t have a spotless record either. Uber’s self-driving cars have been involved in several crashes in Tempe, including a fender bender in September where Uber was not at fault and a crash last March that was significant enough to warrant the temporary grounding of the company’s autonomous-vehicle program. Its vehicles have also been involved in crashes in San Francisco and in Pittsburgh—although in both of those cases a human driver was in control of the car at the time.

Uber likely has good reason for keeping its progress under wraps. It’s certainly a competitive advantage—the fact that Uber has numerous freight trucks already operating on Arizona roadways was a surprise, and it remains unclear which self-driving-car maker is actually leading the space. With its sporadic updates, Uber may be seeking to hide its progress from competitors. If developments are going well, keeping its status under wraps could help lure competitors into a false sense of comfort. Conversely, keeping its mileage and other stats a secret could hide that Uber isn’t as far along as it would like to be. It could also obscure certain metrics—like the number of times truck drivers have had to manually intervene—until the vehicles have amassed enough miles to make those errors seem like rare exceptions to the norm, rather than a common occurrence.

Wherever Uber is on its path toward fully autonomous vehicles, it should take a few cues from Waymo. Its dedicated safety team needs to publish its own self-driving safety report highlighting its accomplishments, challenges, and relevant notable metrics. Congress is pushing to get more autonomous vehicles on the road, but some lawmakers are unconvinced that the technology is ready for our cities and highways—even though, clearly, it’s already there. Transparency is needed to build trust between citizens sharing the roads with these vehicles, those hopping into the back seat one day, and those formulating regulations around their operation. Without mutual trust, these self-driving efforts may as well be dead on arrival.