Oh Look, More Evidence Humans Shouldn't Be Driving

The evidence that human driving isn't the hottest idea keeps coming. This week, we've got two studies to add to the pile.

Saving lives is the best argument for cars that drive themselves. Road crashes killed 1.24 million people worldwide in 2010. Here in the US, they take more than 30,000 lives each year. And more than 90 percent of all crashes can be blamed on human error.

The great promise of autonomous technology is that it will make getting around a whole hell of a lot safer. Robo-cars won't get sleepy, angry, drunk, or distracted. That gives them a massive advantage over us carbon-based lifeforms, which is why automakers will begin offering cars with at least some autonomous tech within three to five years. Yes, there are problems to be solved. The companies pushing this transition have to perfect their hardware and software, and deal with messy regulations and insurance questions.

Then, they have to convince consumers to surrender control. In that battle, at least, they've got plenty of ammunition. There is mounting evidence that letting humans do the driving isn't the hottest idea, and this week saw two more studies they can add to the pile.

On Tuesday, AT&T released research that underscores just how terrible we are at keeping our eyes on the road---a fundamental prerequisite for good driving. We text. We read email. We check Facebook, Twitter, and Instagram. We take selfies, shoot video, and video chat. And "we" means most of us. Distracted driving is the cause of roughly 10 percent of fatal car crashes in the US, according to the DOT.

The AT&T findings are based on a phone survey of more than 2,000 people in the US, aged 16 to 65, who use their smartphone and drive at least once a day. The service provider (which did this research to accompany its "It Can Wait" campaign against distracted driving) doesn't break the numbers down in detail, but says you can't blame phone-addicted millennials and rookies here.

Some more activities we engage in while driving, with the percent of respondents who readily admitted doing them:

  • Text (61 percent)
  • Email (33 percent)
  • Surf the net (28 percent)
  • Facebook (27 percent)
  • Snap a selfie/photo (17 percent)
  • Twitter (14 percent)
  • Instagram (14 percent)
  • Shoot a video (12 percent)
  • Snapchat (11 percent)
  • Video chat (10 percent)

The second study confirms that, just as humans are bad drivers, robots are good drivers. Thatcham Research, a British non-profit funded by insurance companies, drew this conclusion based on accident claims for more than 7,000 VW Golf cars.

All of the cars had adaptive cruise control, which adjusts speed and maintains a safe following distance on the highway. The system also had the ability to apply the brake as needed to avoid a collision. Such features are part of a gradual shift toward autonomy in vehicles take more and more control.

Over the course of one year, third party injury claims on the Golf were 45 percent lower than the equivalent "small family car control group," which included cars like the Ford Fiesta, Toyota Auris, Peugeot 208 and Audi A3. Third party claims are those made by an injured person against the other driver's insurance company.

In other words, people with cars that took over more of the driving hurt fewer people. That's the kind of evidence that will help win over consumers. As long as they're not too busy video chatting to take note.