The National Highway Traffic Safety Administration escalated its probe into Tesla's Full Self-Driving system, citing concerns about its ability to handle poor visibility conditions.
Federal safety regulators are sharpening their focus on Tesla's automated driving-assistance system after raising concerns about the technology's ability to handle poor roadway conditions.
The National Highway Traffic Safety Administration said Thursday that it was escalating a probe of the system, known as Full Self-Driving (Supervised), which controls driving and steering functions but requires drivers to continuously monitor the road.
NHTSA identified several crashes, including one fatal, where Full Self-Driving failed to alert drivers appropriately about reduced-visibility conditions, such as sun glare, fog or airborne dust. The probe originally began in 2024 to examine whether the system can appropriately respond in those scenarios.
NHTSA said Thursday it was escalating the investigation to an "engineering analysis," a more involved examination of a potential safety defect that could result in a recall campaign or other enforcement action.
Tesla didn't immediately respond to a request for comment.
The probe represents the latest effort by NHTSA to examine Full Self-Driving. It also comes one month before Tesla is expected to launch production of a driverless robotaxi called the Cybercab, without traditional controls like a steering wheel or pedals. The vehicle is designed to be driven autonomously by the Full Self-Driving system. Tesla has said the Cybercab would be used as an autonomous taxi and sold to the general public, and it has described the growth of fully autonomous vehicles as key to its shift beyond being a traditional carmaker into an artificial-intelligence and robotics powerhouse.
NHTSA, the top U.S. auto-safety regulator, has spent years examining Tesla's automated-driving technology. The agency is also scrutinizing similar systems from Tesla's competitors.
So-called automated driver-assistance systems like Full Self-Driving require drivers to remain attentive to the road at all times and resume control of the vehicle if prompted. NHTSA has spent years examining whether drivers are sufficiently warned if the system disengages.
For Tesla, NHTSA is particularly focused on the company's unique design of Full Self-Driving, which relies on cameras and a "vision-based" setup. Unlike fully autonomous robotaxis from companies like Alphabet-owned Waymo, or similar driver-assistance systems from burgeoning competitors, Tesla doesn't use radar or lidar—a light-based system that allows cars to "see" a fuller picture of the world around them.
When Tesla transitioned away from using both cameras and radars to the vision-based system in mid-2021, the automaker developed and implemented a "degradation detection system" that it deployed through a software update to existing and new vehicles, NHTSA said.
Tesla began developing an update to that system in June 2024, the regulator said, but as of now, NHTSA doesn't know whether the vehicles received the system.
In the filing Thursday, NHTSA said that incident data the agency reviewed raises concerns that the degradation detection system, both as originally deployed and later updated, fails to detect poor visibility conditions and warn the driver appropriately.
In crashes that NHTSA reviewed, the system didn't detect "common" roadway conditions that impaired visibility and didn't provide alerts to the driver until immediately before a collision occurred. In each of these cases, Full Self-Driving "also lost track of or never detected" a vehicle in front of the system's path.
NHTSA's probe covers about 3.2 million vehicles and includes all models equipped with Full Self-Driving, such as certain versions of Model 3, Y, X and S, as well as the Cybertruck. NHTSA said it would gather additional information about the degradation detection system and analyze a half-dozen potentially related incidents.