The recent car accident in Tempe, Arizona that took the life of 49-year-old Elaine Herzberg has seemingly slammed the brakes on the driverless future racing toward us. While social media made cracks about the beginning of the robot uprising, old-timey media-stoked fears of autonomous technology.
In the aftermath, Uber has put their program on hold, as have Toyota and tech-startup NuTonomy, though others are still moving forward. In fact, Google spinoff Waymo still plans to offer self-driving vans in Phoenix later this year.
I get the visceral reaction. Tech companies like Facebook are increasingly untrustworthy these days and we’ve been afraid of artificial intelligence since long before Terminator. One of my favourite books is Daniel H. Wilson’s similarly themed Robopocalypse, a 2011 novel about the Great Robot War which begins, naturally, with smart cars mowing down pedestrians on purpose.
The Tempe incident was not on purpose, of course. Herzberg was jaywalking her bike across a dark road at night and Uber’s anti-collision software failed to prevent the collision, as did the distracted emergency human in the driver’s seat. The result was Herzberg becoming the very first person in history to be killed by a driverless car. (But not the first killed in one, which happened to a Tesla operator in 2016 when his autopilot vehicle hit a truck.)
It was an unquestionable tragedy. But the only unique thing about it was the driver (or lack thereof). About 40,000 Americans have killed in car accidents annually, as are around 2000 Canadians and 1.4 million people worldwide. In North America, motor vehicle accidents are the leading cause of death for young people from age 1 to 34. More than 90 percent of these deaths are a result of (human) driver error. But we largely let all that slide.
Anti-vaxxers are so scared of adverse vaccine reactions, which are rare enough to be unmeasurable in data studies, that they’ll risk their kids getting measles or whooping cough. But they will also drive those same kids around town in their far more dangerous cars.
This is called cognitive dissonance — unlikely threats are far more likely to resonate in our minds than everyday ones. It’s also why people are so afraid of terrorism, yet not bothered by killer cars.
But we should have no misconceptions about the safety of human drivers. As Toronto mayor John Tory told the Globe & Mail in 2016 about the news that the number of pedestrians killed by cars had risen 15 percent in the previous five years — with someone dying every ten days and being hit every four hours.
“If you compare it to how much attention is paid each year to the number of people who are killed by homicide, or a number of other things, it has received less attention than it should, especially given the magnitude of the number,” Tory said.
And yet that year Toronto hit another record with 243 pedestrians, and a cyclist was killed. I doubt a robot driver would have nearly sideswiped me on my bike this week while speeding to pass a streetcar or knocked me right off my bike last year with a trailer that was slightly wider than the truck pulling it. I was down in Austin, Texas a few years back when another not-robot driver crashed into a crowd during the SXSW music festival, killing four concertgoers.
And while there were no deaths or serious injuries, I was in two serious car accidents myself as a teenager. One involved a speeding car that hit mine as I was left-turning onto a highway, in the other I failed to see a stop sign and my front driver’s side got hit by a GMC truck. I escaped with a cut face and a broken leg. My older sister’s Tercel was totalled.
Humans are easily distracted and often tired, angry, frustrated or intoxicated. They like to speed. They miss seeing things. They make mistakes, and those mistakes cost lives. Yes, we do at least know who to blame in those instances whereas it’s hard to send a Volvo XC90 SUV to prison for vehicular manslaughter.
There is no question that the technology needs to be improved before these cars are ready to join us on our roads. A New York Times investigation found that internal Uber documents revealed their “cars were having trouble driving through construction zones and next to tall vehicles, like big rigs. And Uber’s human drivers had to intervene far more frequently than the drivers of competing autonomous car projects.” The Times also found that Uber was pushing their limits to “live up to a goal to offer a driverless car service by the end of the year and to impress top executives.”
Should Uber get out of this market? Probably. We certainly can’t let corporations risks our lives to meet their bottom lines. But at the same time, humans are even worse drivers and they don’t deserve a free pass, either.
“We are so numb and tolerant of the crashes that occur by the thousands all around us every year,” RAND Corporation researcher David Groves told Wired last year. “But the first autonomous vehicle crash is going to be extremely novel.”
“The public will flip the first time one single person dies in a self-driving car accident,” accurately predicted the article’s author, “even if thousands of others have been ‘saved’ by non-distracted, non-drunk robo-cars.”
And that’s the rub. American researchers have found that even if autonomous cars are only 10 percent better than the average driver — meaning they still cause a fair number of deaths — they could still save as many as 3,000 lives a year. A 2015 study predicted that by mid-century autonomous vehicles would have the effect in the U.S. of “reducing accidents by up to 90 percent.” But last year, long before Tempe, Pew Research Center found 56 percent of Americans “would not want to ride in a driverless vehicle if given the opportunity,” citing safety concerns about their robot drivers.
Maybe I’m less concerned because I grew up on the west coast where the Skytrain public transit system has been driven by robots since the mid-1980s. Yes, there are less potential hazards on a raised or buried track than on the roads, but there are also fewer deaths on Skytrain than on the human-driven Toronto subway system. (In both cases, most are suicide-related.) There’s certainly been nothing involving driverless Vancouver trains like the Toronto subway crash in 1995 that left three dead and 30 hospitalized when a new driver ran three red lights and smashed into a stopped train ahead. But Torontonians didn’t abandon the subway, they improved safety measures.
Which brings us back to Tempe, an accident that experts say would have still happened with a human driver, albeit without the press coverage. Because we accept that that happens every day. To a degree, it’s like the U.S. banning water bottles on planes because of one failed terror attack but doing nothing about controlling guns which kill around 40,000 people annually. I have no problem with heightening security because of the former but it’s hardly a threat like the latter.
We hear a lot about the war on cars, but given the casualty count it’s more like cars have declared war on us. So let’s iterate and regulate rather than abandon autonomous vehicles because human-driven cars are much more likely to leave you dead.