Critics Want Tesla to Justify Safety Claims for Autopilot

Safety group are pushing Tesla Inc. to justify its claim that its semi-autonomous-driving system makes its electric cars 40% safer, notes Automotive News.

Tesla has used the figure to tout the benefits of its Autopilot self-driving system, which can control a vehicle automatically under certain conditions. A federal court lawsuit filed last year is trying to force the release of the underlying data used to calculate the number.

The estimate is based on an analysis by the National Highway Traffic Safety Administration. NHTSA studied crashes of 2014-2016 model Tesla Model S sedans and Model X crossovers in which airbags deployed. The calculations use mileage data supplied by Tesla.

NHTSA calculated crash rates before and after Tesla introduced Autosteer, the robotic steering component of the Autopilot system. The agency says the crash rate per million miles driven dropped nearly 40% to 0.8 from 1.3 after Autosteer software was loaded into the cars, regardless of whether the system was turned on.

Tesla, claiming competitive concerns, asked NHTSA to keep the underlying data confidential, and the agency has complied.

Skeptical researchers and safety experts tell AN that Tesla is inaccurately extrapolating NHTSA’s findings. For one thing, it isn’t clear whether the mileage figures Tesla provided are sufficient to generate statistically significant results. It also isn’t clear how the Autopilot system can be credited with reducing crashes even when it isn’t turned on.

The Insurance Institute for Highway Safety estimates a 13% reduction in the frequency of insurance claims involving Model S EVs equipped with Autopilot compared to Model S cars without the system. IIHS’s calculations involve a broader category than insurance claims that also involve airbag deployments.

Even so, the institute says it lacks sufficient data to attribute the difference it found to any specific technology aboard the cars.