Some people are happy about the slow, steady arrival of self-driving cars. Others are terrified. But in one new survey, people on both sides of the divide say that they're unlikely to let autonomous cars do all the driving, despite humankind's demonstrated penchant for accidents.
The survey was conducted by the folks at Insurance.com, who polled 2,000 licensed drivers in June 2014. Among the more notable findings:
- When asked whether they'd buy an autonomous car, 22.4 percent of consumers said that they'd be very likely to do so, but 24.5 percent say that they'd never invest in one.
- When asked whether they'd buy an autonomous car if it meant an 80 percent reduction in their insurance premium, the figures changed dramatically: 37.6 expressed interest, while just 13.7 percent said no way, no how.
- When asked how often they'd let an autonomous car do the driving, though, just 31 percent said that they'd let the car's computer take over whenever possible.
- When asked whether they'd let an autonomous car drive their kid to school, only 23.8 said yes.
- And the kicker: 61 percent of those surveyed said that they would make better driving decisions than a computer.
As much as we'd love to believe in our own infallibility, we know for a fact that computers drive better than we can. Maybe not today, but very, very soon, when the software is tidied up, autonomous cars will make far better drivers.
How can we -- people who ostensibly love driving -- say such a thing? Because it's true. We'll still be better at taking selfies and making movies and flirting and lots of other human things, but when it comes to driving, we can see the writing on the wall.
Consider this: in 2012, there were 33,561 traffic fatalities on U.S. roads. Of those, 10,322 (30.8 percent) were caused by drunk drivers. Another 10,219 (30.4 percent) were caused by speeding. And 3,328 (10 percent) were caused by distracted drivers. Even if we were to account for some overlap between the three (e.g. speeding texters), alcohol, speeding, and distractions led to the vast majority of traffic deaths.
Obviously, autonomous cars don't get hammered, they don't get distracted by text messages, and it's doubtful that their software will allow them to speed (with the possible exception of road trains). So, if the cars involved in the 30,800 collisions that led to those 33,561 deaths had been autonomous, it's fair to say that the bulk of those fatalities could've been avoided.
But that's not all: human error accounts for up to 95 percent of all traffic accidents. So, of the 5,615,000 fender-benders and other non-fatal accidents that took place in the U.S. in 2012, it's fair to say that most -- along with a goodly portion of the 2,362,000 injuries they caused -- could've been avoided if computers had done the driving.
The odds of a computer besting you behind the wheel only improve when vehicle-to-vehicle technology becomes widespread. V2V lets cars "talk" to each other, as well as stop lights and other elements of infrastructure, "seeing" and stopping accidents long before humans could react.
Admittedly, there are some accidents that autonomous vehicles won't be able to dodge -- not cleanly, anyway. Natural disasters. Runaway dogs. Distracted walkers. And of course, once cars are networked, there's the possibility of hacking (which the National Highway Traffic Safety Administration wants to nip in the bud).
But those are extreme situations. Looking at this with clear eyes and a level head, we have to consider the odds -- and the odds are definitely in the computer's favor. Any discussion of "what about this, what about that" is a distraction (as we learned during the Cold War), and distractions are just another human weakness.