Tesla’s Full Self-Driving software under investigation by federal safety regulator

The top U.S. automotive safety regulator has opened a new investigation into Tesla’s so-called “Full Self-Driving (Supervised)” software after four reported crashes in low-visibility situations — including one where a pedestrian was killed.

The National Highway Traffic Safety Administration’s (NHTSA) Office of Defects Investigation (ODI) announced Friday that it is probing the driver assistance system to find out whether it can “detect and respond appropriately to reduced roadway visibility conditions,” such as “sun glare, fog, or airborne dust.” The agency also wants to know if other crashes have occurred in these conditions beyond the ones that were reported.

The investigation comes just one week after Tesla CEO Elon Musk revealed the prototype of his company’s “CyberCab,” a two-seater car that he said is supposed to serve as a robotaxi, following years of unmet promises. Musk also claimed at the event that Tesla’s Model 3 sedan and Model Y SUV would be able to operate without supervision in California and Texas at some point in 2025, though he did not offer any details about how that will happen.

In April, NHTSA closed a nearly three-year probe into Autopilot, the less-capable driver assistance software that Tesla offers, after investigating almost 500 crashes where the system was active. The agency found 13 of those crashes were fatal. At the same time that it closed that probe, NHTSA opened a new investigation into the recall fix that Tesla had issued to address problems with Autopilot.

Tesla’s software also faces other legal threats. The Department of Justice is investigating the claims Tesla has made about its driver-assistance features, and the California Department of Motor Vehicles has accused Tesla of inflating the software’s capabilities.