As interest in self-driving cars accelerates, so have questions about their safety and reliability. The idea is that self-driving cars will be safer because they will make far fewer mistakes. But can people trust computer algorithms to guide them safely through incredibly complex and ever-changing road conditions? There have been reports of some driverless cars getting into crashes, but how do they compare with the real world of human-operated ones?
Researchers at the University of Michigan’s Transportation Research Institute set out to answer that question. They compared the safety record for all vehicles on the road in 2013 with the safety record of driverless cars — made by Google, Delphi Automotive and Volkswagen’s Audi division — on the road from 2012 to 2015.
They found that per mile driven, driverless cars were actually far more likely to get in a crash than human-operated ones. While conventional vehicles had 4.1 crashes per million miles of travel, self-driving cars had 9.1. And while the injury rate for human-operated cars was 1.02 per million miles driven, for driverless cars it was 3.29.
There’s a caveat, of course:
Those are large spreads. But the authors caution that driverless cars have very limited total mileage so far — just 1.2 million miles — compared with 3 trillion miles for conventional cars. As a result, they can’t rule out “the possibility that the actual rates for self-driving vehicles are lower than for conventional vehicles.” They also note that the severity of the crashes that involved driverless cars tended to be lower. The biggest difference that the research found, however, is fault. So far, driverless cars were to blame for zero of the crashes.
In other words, the jury is still out. But speaking as a red-blooded American male, I hate the idea of driverless cars, partly because they’re unmanly and mostly because they’ll soon enough be mandated by the nanny-state fascists who already control too much of America. The hell with them.