Tesla made it into the history books last month, and for all the wrong reasons. It earned the dubious distinction of being the first automaker to record a fatality in a semi-autonomous car.
A Tesla owner was killed when the car’s Autopilot system could not detect that a semi-truck was blocking the road and ran the car right into it. That set off a media frenzy on the dangers of autonomous technology. But this accident really should not have surprised anyone. And it should not deter the auto industry from pressing forward with autonomous cars.
Tesla will bear the responsibility for this accident. Its system failed spectacularly. Tesla can argue its cars warn drivers to remain alert and to keep their hands on the steering wheel. But it treats that legal disclaimer the same way the rest of us treat the fine print in a television commercial – something to be ignored. Last year Tesla’s CEO, Elon Musk, bragged that owners with Autopilot basically would be able to drive from San Francisco to Seattle without the driver doing anything.
Mercedes and Volvo also sell semi-autonomous systems that operate much like Tesla’s. But their systems force drivers to keep their hands on the wheel. If they don’t, a warning icon appears on the instrument panel instructing the drivers to grasp the wheel. And if they still don’t, the system automatically disengages within seconds and the car starts to slow down.
Even though this technology is in its infancy, it promises dramatic improvements in automotive safety. We know that over 90% of all accidents are caused by human error, so taking the driver out of the equation will result in dramatic improvements in safety. No, autonomous cars won’t be perfect. Yes, people still will get killed in autonomous cars. But it will be far fewer than are killed today.
The statistics are chilling. Yesterday, more than 100 people were killed in traffic accidents in the U.S. Today, another 100 will be killed. Tomorrow, another 100. The sad fact of the matter is that over 100 people are killed every day of the year. And over 6,000 are hurt badly enough that they need to go to the hospital. That’s 6,000 a day. It almost seems a little silly to get hung up on one fatality in a semi-autonomous car when this technology offers the potential to eliminate 80% to 90% of all accidents.
Back in 1986, when the Challenger space shuttle exploded and killed all the astronauts on board, there was a similar amount of national hand-wringing about safety. A panel of experts that NASA brought together quickly concluded cold temperatures on launch morning caused a gasket to fail which led to the disaster. Yet NASA shut the entire shuttle program down for nearly three years to do more studies.
I remember a television reporter at the time asking Chuck Yeager, the famous war hero and test pilot, what he thought NASA should do. “I think they should wait for a warm morning,” he drawled, “and shoot off the next one.” Three years later that’s basically what NASA ended up doing.
That’s the attitude we need in developing autonomous cars. Shoot off the next one. The benefits are too big to ignore, and we can measure any delays with a body count. And that’s why I’ve said and will keep saying: We lose by waiting.