When Waymo, Alphabet's self-driving car division that started in 2009 as a Google initiative, decided to offer an autopilot "out" for passengers—meaning they could take over the vehicle's controls if a dangerous situation arose—it seemed like a practical and needed feature. But the company's CEO now says they eventually opted not to offer such a safety feature, because it appeared to not be very safe at all, per Reuters. Waymo chief John Krafcik says what the self-driving team saw during 2013 test runs, which involved Google employees riding in the vehicles with video cameras installed to survey the scene, was "pretty scary." The videos showed that human passengers lost track of what was going on, instead checking their phones, doing their makeup, and even dozing off—making it difficult for them to jump behind the wheel at a second's notice.
"It's hard [for drivers] to take over because they have lost contextual awareness," Krafcik said. Instead, Waymo passengers in a pilot program in Phoenix are now given limited controls in their Chrysler Pacifica minivans via four buttons only, per Wired: a "Help" button, a button to lock and unlock the doors, a "Start Ride" button, and a "Pull Over Here" button, which allows a passenger to hop out at any point during the ride or pick up another passenger. "Our technology takes care of all of the driving, allowing passengers to stay passengers," a company report said earlier this month. The Verge mentions other Waymo "complications and unanswered questions," which the company is said to be "furiously working to address" in its quest for full self-driving autonomy. "Our goal … is to bring this technology to the public—to the world," Krafcik says. (This Silicon Valley job pays a "startling" salary.)