Let’s say you’re driving down Main Street and your brakes give out. As the terror hits, a gaggle of children spills out into the road. Do you A) swerve into Keith’s Frozen Yogurt Emporium, killing yourself, covering your car in toppings, and sparing the kids or B) assume they’re the Children of the Corn and just power through, killing them and saving your own life? Any decent human would choose the former, of course, because even murderous kiddie farmers have rights.
But would a self-driving car make the right choice? Maybe yes. But even if it does, by programming a machine to save children, you’re also programming it to kill the driver. This is known as the trolley problem (it’s older than self-driving cars, you see), and it illustrates a strange truth: Not only will robocars fail to completely eliminate traffic deaths, but on very, very rare occasions, they’ll be choosing who to sacrifice—all to make the roads of tomorrow a far safer place.
Read the full article here – https://www.wired.com/2017/03/make-us-safer-robocars-will-sometimes-kill