Tesla’s Autopilot Vindicated With 40% Drop in Crashes
Tesla’s driver-assistance features, known collectively as Autopilot, have been vindicated.
The U.S. Department of Transportation has closed its investigation into a May 7 collision with a tractor-trailer that killed a driver using Autopilot. The agency found no indication of a safety problem with it. In fact, the evidence provided by Tesla included crucial data that’s been missing from the safety debate surrounding automated cars: crash rates.
Tesla is in a unique position to determine the precise impact of Autopilot on crash rates, more so than any other car manufacturer. That’s because while all Tesla vehicles come with the hardware necessary for Autopilot, you need a software upgrade that costs thousands of dollars to make it work. Since buyers can add Autopilot features after purchase, this provides a perfect before-and-after comparison.
It turns out that, according to the data Tesla gave investigators, installing Autopilot prevents crashes—by an astonishing 40 percent. The chart below comes from the report NHTSA issued while concluding its investigation. Approximately one-third of the mileage on the cars was logged before the upgrade to Autosteer (the most controversial component of the driving suite), while the remaining miles were accrued after installation.
Tesla’s reputation took a hit after it disclosed the death that triggered the federal probe. It’s no exaggeration to say it changed the way people think about self-driving cars, despite the thousands of deaths annually tied to ordinary vehicles. Consumer Reports went so far as to call on Tesla to revoke the features until changes were made. What Tesla had initially designed to be a safety feature had suddenly become a liability.
The only statistic that Tesla offered in its public defense at the time was flimsy: Teslas had been driven on Autopilot for 130 million miles before the first fatal crash, compared with a national U.S. rate of 94 million miles per fatality. With only one death, this was hardly good statistical evidence, especially when one considers that the Tesla Model S has the highest safety rating to begin with and Autopilot is recommended only for highway driving.
Now—thanks to an investigation that initially hurt the company—there is finally some real data, and it’s good news for Tesla.
“Tesla is not under any active investigation after today,” said NHTSA spokesman Bryan Thomas.
A separate investigation into the fatal accident by the National Transportation Safety Board is ongoing and will present conclusions by early summer. The NTSB is an independent agency that has no regulatory power.
Meanwhile Autopilot continues to evolve. Tesla rolled out a new version of its software in November, known as Tesla 8.0. The update requires drivers to touch the steering wheel more frequently and increases Autopilot's reliance on radar, in addition to cameras and ultrasonic sensors. Tesla Chief Executive Officer Elon Musk said 8.0 would have been able to detect the truck that was involved in the fatal broadside accident.
In October, the company began shipping new Autopilot hardware with enhanced sensors that it says will eventually enable fully autonomous driving. Every car now ships with eight cameras and a dozen sensors to give 360-degree visibility. The company is rolling out new features that make use of the sensor suite in regular over-the-air updates.
As the software matures to match the new hardware, Musk said on Thursday via a Tweet, Tesla is targeting a 90 percent reduction in car crashes.