Tesla Pumps the Brakes on Autopilot, Gears Up for Fully Autonomous Cars

Tesla made a brief announcement about their current slate of electric vehicles last night, which seems to be something of a temporary step back from Tesla’s semi-autonomous Autopilot feature. Elon Musk announced that starting now, every Tesla vehicle produced will be equipped with the hardware necessary for fully autonomous driving, with the stipulation that Autopilot features will temporarily be disabled on those new cars.

The announcement covers the Model S, Model X, and Model 3, which will all be outfitted with eight cameras that give each car a full 360-degree view of its surroundings for up to 250 meters. Each car will also have twelve ultrasonic sensors for detecting the presence of obstacles on all sides. Radar at the front of the car also scans for obstacles and can work through inclement weather, like rain, fog, or dust storms. Meanwhile, the in-car computer is 40 times more powerful and runs Tesla’s latest software, which processes and makes sense of all the data coming from those sensors.

However, the Tesla press release notes that those new vehicles will, “temporarily lack certain features currently available on Teslas with first-generation Autopilot hardware.” That includes automatic emergency braking, collision warning, lane holding, and active cruise control. Tesla plans to activate those features in a future update, but they didn’t mention a time frame.

Those features were at the center of an incident in June, when a driver was killed while Autopilot was activated in his Model S. More Autopilot accidents were reported after that, leading to Germany’s transport ministry to instruct Tesla to not call the feature “Autopilot,” claiming the name is misleading. Tesla owners had been instructed to remain alert with both hands on the wheels while using the feature, but a combination of a lack of driver attention and software deficiencies (in the June incident, Autopilot failed to recognize the trailer of a semi from the side as an obstacle because it was so high off the ground) has soured perception of Autopilot.

Last night’s announcement was an admission of sorts that introducing partial autonomy so quickly was a mistake. Although Tesla was consistently up front about the limitations of the feature and the responsibilities of the drivers (who had to sign a waiver), calling it Autopilot is certainly misleading — at most, it can only keep the car going straight in a lane and stop when it detects obstacles, and even then, the software still has a ways to go. The promise of autonomous cars is safety, with the implicit guarantee that self-driving cars will be safer than human drivers. Partial autonomy retains the human element, making Autopilot more like a reckless beta program than a safety feature.

It’s worth noting that today’s announcement that Tesla’s cars now have the hardware for autonomous driving doesn’t come as a huge surprise. Equipping cars with sufficient hardware is, relatively speaking, the easy part — a lot of work needs to be done to not only improve the drive software, but to establish connected infrastructure within cities to ensure that autonomous driving is as safe as it could and should be. Here’s hoping Tesla gets there, though — autonomous cars from Google and Tesla alike have had very low accident rates during testing, suggesting that they could indeed save millions of lives per year if their use becomes widespread.