Tesla reported most automated system-related crashes, NHTSA says

The news comes from the first public release of data on crashes involving level 2 automated driving systems.

“These technologies hold great promise to improve safety, but we need to understand how these vehicles are performing in real-world situations,” NHTSA Administrator Steve Cliff told reporters ahead of the release of the data. (Credit: Angel Garcia/Bloomberg)

(Bloomberg) — Tesla Inc. reported the vast majority of crashes involving automated driver-assist systems that have been disclosed to the U.S.  National Highway Traffic Safety Administration (NHTSA), according to new figures from the regulator that it says are too limited to draw any safety conclusions.

Tesla accounted for 273 of 367 such crashes reported by 12 auto companies to NHTSA between July 2021 and May 15 of this year, the agency said. Across all reported accidents, serious injury or death occurred in 11 of the 98 collisions for which severity data was collected. A further 294 incidents lacked information about harm to vehicle occupants.

Behind Tesla was Honda Motor Co. with 90 reported crashes, while Subaru Corp. disclosed 10 collisions. All remaining companies including General Motors Co., Ford Motor Co. and Toyota Motor Corp. cited five or fewer collisions.

The figures were included in the first public release of data collected about crashes involving so-called level 2 automated driving systems, a broad look at how technologies that automate some of the driving task in certain situations but need constant monitoring by drivers have performed in real-world settings.

Data collection mandate

The reports were collected under a June 2021 order demanding carmakers and technology companies report the incidents. Its release comes as safety advocates in Washington call for more action from regulators and lawmakers to set firmer rules for so-called self-driving cars, as technologies such as Tesla’s driver-assist features gain popularity with customers.

NHTSA also reported crash figures for vehicles equipped with automated systems designed to drive a vehicle without a driver that are being developed by companies such as Cruise LLC, the robotaxi startup controlled by GM. Some 25 companies reported a total of 130 crashes over the same period, with Alphabet Inc.’s Waymo self-driving car unit reporting 62, followed by Transdev Alternative Services with 34. Cruise disclosed 23 collisions.

Data just a ‘small piece’ of the picture

Advocates for self-driving cars and driver-assist systems point to the potential to reduce the millions of crashes that occur in the U.S. each year and curb traffic road deaths, which have recently spiked and fueled worries among safety advocates and policymakers. More than 9 million vehicles of all types were involved in police-reported crashes in 2020, according to NHTSA.

“These technologies hold great promise to improve safety, but we need to understand how these vehicles are performing in real-world situations,” NHTSA Administrator Steve Cliff told reporters ahead of the release of the data.

At the same time, NHTSA said the figures shouldn’t be used to make safety conclusions. The data lacked contextual information needed to establish a rate of incidents, such as the number of vehicles in each manufacturer’s fleet that were equipped with driver-assist systems, how often drivers use them or the number of miles driven. Crash reports may also be duplicative, it said.

The Alliance for Automotive Innovation, the industry’s main trade association, underlined NHTSA’s caution, saying that broad conclusions about the safety of these systems shouldn’t be drawn from the new information.

“This is important data, but it’s only a small piece of the overall picture,” John Bozzella, the group’s chief executive officer, said in a statement. “There’s more work to do.”

The new data comes soon after the agency escalated an investigation into whether Tesla’s Autopilot system is defective. It opened the probe into a possible defect of the electric-car maker’s partially automated driver-assistance feature in August 2021, when it began looking into how the system handles crash scenes following a dozen collisions with first-responder and other vehicles.

Related: Autonomous trucks would haul new risks into commercial lines

In spite of its limitations, NHTSA said the data will help it better understand how the systems are performing in the field, potential risks to pedestrians and other vulnerable road users, rule-making efforts and enforcement actions. The agency will also update the data on its website monthly, it said.

“This will help our investigators quickly identify potential defect trends that could emerge, some of which will warrant further explore exploration,” Cliff said.

In December, NHTSA launched an evaluation after reports of Tesla car occupants playing video games on front-center touch screens. The carmaker told the agency it would work on a software update to lock the feature when vehicles are in motion.

Tesla has marketed driver-assistance features using the names Autopilot and Full Self-Driving that still require drivers to keep their hands on the wheel. The company has drawn criticism from the likes of the National Transportation Safety Board, former NHTSA leaders and members of Congress over issues including how it has branded the systems and whether it does enough to safeguard against inattentiveness and misuse.

Related:

Copyright 2024 Bloomberg. All rights reserved. This material may not be published, broadcast, rewritten, or redistributed.