As Google is working on rolling out the latest version of its self-driving cars for road tests this summer, the company has also revealed that its vehicles have been involved in a total of 13 accidents since the program began. That number was accrued over nearly two million miles of testing in six years, and most importantly, all of the accidents have been the fault of other drivers, not Google’s autonomous cars.
Considering how many miles the cars have traveled and how many years they’ve been on the road in some form or another, it’s almost surprising that there haven’t been more incidents. It’s definitely surprising none of the accidents have been the fault of Google’s autonomous technology. Despite people’s assumptions that computers can’t make the necessary decisions required to avoid accidents, it appears that at least for now, that has yet to be the case.
News of Google’s driverless cars being involved in several accidents spread quickly, and a number of consumer groups, especially one called Consumer Watchdog, raised concern. Part of this concern is because the newest generation of cars that Google is testing have no driver inputs whatsoever and operate entirely autonomously, but there were also concerns that Google hasn’t been transparent enough about the accidents its cars are involved in.
“Google is dribbling out bits information in the hope to silence legitimate calls for full transparency. They are testing on public roads and the public has a right to know exactly what happened when something goes wrong,” said John Simpson, Consumer Watchdog’s Privacy Project director.
As Consumer Watchdog points out, despite Google’s essentially-perfect track record so far, there are still significant shortcomings in autonomous driving technology. Among the several that Consumer Watchdog lists, two of the most significant ones are how autonomous cars handle inclement weather and changing traffic conditions.
Sensors still don’t work well in the rain, and they perform even more poorly in the snow. Not allowing a passenger to take control and drive the car in those situations would make self-driving cars incredibly dangerous. As for changing traffic conditions, anything from road construction to a new traffic signal will still seriously confuse a driverless car that can’t use human judgment to make the necessary decisions.
At the same time, despite these and other concerns, Google’s cars have yet to be actually cause any accidents while in autonomous mode. Every accident Google’s cars have been involved in has been the fault of another driver, including one incident involving a Google employee operating the car manually. Of those 13 accidents, seven of them were rear-end crashes. Whether a car has a driver at the wheel or a computer, how is someone stopped at a traffic light supposed to keep another driver from rear-ending them?
Despite the initial shock that Google’s cars had been involved in accidents, people who think this news reflects poorly on the company are mistaken. It’s proof that as long as everything is working properly, riding in a computer-driven car is safer than riding in a human-driven car. Computers don’t stop looking at the road to send text messages, read books, or eat burritos. Those are, however, things that human drivers love to do.
In general, it’s a good idea for Google to be upfront about incidents that its self-driving cars are involved in, and there are quite a few advances in technology that will need to happen before fully-autonomous vehicles can be considered fit for sale to the public. That said, until one of Google’s cars is actually found to be at fault for a wreck while operating autonomously, no amount of extra information that Google shares is going to make the public any safer.