Back in December, as Uber transferred its controversial self-driving pilot program from California to Arizona, a police officer in Phoenix offered a rather innocuous comment about what lies ahead: “Now you’ve introduced these situations where you’ve got manned vehicles and unmanned vehicles,” he said.
One of those situations—a three-way crash in the city of Tempe involving a self-driving Uber that, police say, isn’t at fault—transpired over the weekend, and it illustrates the potential problems of mixing semi-autonomous and manually-driven vehicles on the road.
There were no serious injuries reported, but the crash shows the inherent risk of testing self-driving cars on the road with humans. There will be accidents; hell, California keeps a list of them. But at a time when lawmakers everywhere are still working out the kinks on how to regulate the fledgling technology, a simple crash could stoke the fear of those who just aren’t ready to see a robot car on the road next to them.
The disproportionate amount of coverage given to a non-fatal crash involving a semi-autonomous Uber, especially when more than 40,000 people died last year in regular car accidents, says enough about how sensitive people are toward the emerging technology.
Those concerns aren’t immune to self-driving boosters. Take Arizona Governor Doug Ducey. After the report quoting the Phoenix officer ran, Phoenix city manager Ed Zuercher shot an email over to the city’s police chief and said, while the story “isn’t bad,” they ought to be “careful about getting pulled into this matter from a [police department] perspective.”
“The mayor and governor are both sensitive about this subject,” Zuercher wrote in the email, obtained by Jalopnik under a public records request.
Ducey is an impassioned supporter of Uber—when the company announced its decision to move to Arizona, he gushingly said, “California may not want you, but we do”—but in an environment where universal regulations overseeing autonomous vehicles are negligible, a high-profile crash is something that could derail the momentum for producing fully autonomous cars. (Ford’s CEO said as much last month.)
Ducey’s spokesperson declined to comment on the email without seeing it, but did offer that the governor “expects Uber and other companies in the testing phase to fully comply with all of Arizona’s state and municipal laws, and he also expects law enforcement to fully enforce those laws.”
Uber took an extremely cautious approach, as well. The company announced on Saturday that it was grounding its self-driving pilot programs in Pittsburgh, Arizona and San Francisco, where it now has a proper permit to operate, pending the completion of an investigation into the incident. On Monday, a spokesperson said in a statement that “development operations” in San Francisco have resumed. Vehicles in Arizona and Pittsburgh remain grounded, but Uber says it expects them on the road again soon.
It was a brief, but significant decision. All for what, again, appears to be a fairly innocuous accident. Tempe police say around 6:30 pm. on Friday, a self-driving Uber was heading southbound on a road when a separate vehicle failed to yield while turning left. The vehicles collided, and the self-driving Volvo rolled onto its side.
An Uber driver was behind the wheel, an engineer in the passenger seat; the car was in self-driving mode. There was no passenger and no reported serious injuries. (Tempe police said a full accident report will be released upon its completion.)
This isn’t to object to testing self-driving cars on public roadways; autonomous tech itself has a long way to go as it is: leaked internal reports show Uber’s cars need a human to take the wheel about every mile.
But the crash in Arizona over the weekend just reiterates the reality of the situation: It’s simply going to be a long time before the auto market is overtaken by self-driving vehicles. In the meantime, with a small number of vehicles already on the road in pockets of the U.S., the potential for a crash involving AVs and human-driven vehicles will continue to exist.
A study released by the Governors Highway Safety Association last month urged states to develop education campaigns about AVs, as well as how to share the road with them.
“The research and media attention given to autonomous vehicles often overlooks the safety implications that a mix of driver-operated and autonomous vehicles will bring,” said Dr. James Hedlund, a former senior official with the National Highway Traffic Safety Administration and co-author of the study, in a statement at the time. “Unfortunately, ignoring the driver side of the equation may negate many of the expected safety benefits.”
The association also said states should wait until model laws and regulations are developed at the federal level to encourage more uniformity of regulations, not a patchwork. It’s perhaps at least of some comfort that new U.S. Secretary Elaine Chao says her department is reviewing guidelines offered by the previous administration.
The concern, then, is that mixing both robot cars and humans on the road could scuttle advances with the technology, and the purported safety benefits that come with it. Human error is attributable to more than 90 percent of crashes. It’s still unclear if the Uber driver in the Arizona accident would’ve had time to take the wheel back and hit the brakes in time. (Humans, studies have shown, aren’t particularly good at resuming control of the car if a semi-autonomous vehicle needs it to.)
But the belief of AV proponents is, to drastically cut the number of fatalities, you need to take humans out of the equation. The Uber crash shows how tight of a rope automakers and Silicon Valley are walking to make that a reality.