The U.S.’ high automotive security regulator is notching up its investigation into the efficiency of Tesla’s Full Self-Driving (Supervised) driving software program in low-visibility situations.
The Nationwide Freeway Site visitors Security Administration’s Workplace of Defects Investigation (ODI) said on Thursday that it has upgraded the probe it launched in October 2024 to what’s referred to as an “engineering evaluation,” its highest stage of scrutiny. It’s a step that’s typically required earlier than the company tells an organization to subject a recall.
That is certainly one of two investigations that ODI is working on Tesla’s Full Self-Driving (Supervised) software program. The regulator can also be probing greater than 80 cases wherein Tesla’s driver-assistance software program has violated primary visitors security legal guidelines, like working purple lights. The investigations come as Tesla has spent months making an attempt to get a robotaxi service off the bottom in Austin, Texas.
ODI opened this specific probe after 4 reported crashes in low-visibility conditions, certainly one of which concerned the demise of a pedestrian. The regulator has spent the final year-and-a-half exchanging info with Tesla, and seems to have recognized a handful extra incidents the place the corporate’s driving software program proved inadequate in low-visibility settings.
ODI additionally mentioned on Thursday that it has not gotten all the knowledge it desires from Tesla on this course of. The investigative workplace wrote that, whereas Tesla started “growing an replace” to repair the low-visibility issues in June 2024 — earlier than the probe was even opened — the corporate has nonetheless not advised ODI whether or not that repair was deployed, or which autos acquired it.
ODI additionally believes there may very well be an under-reporting of comparable crashes attributable to knowledge assortment and labeling limitations that Tesla reported to the security company.
“Within the crashes that ODI has reviewed, the system didn’t detect frequent roadway situations that impaired digital camera visibility and/or present alerts when digital camera efficiency had deteriorated till instantly earlier than the crash occurred,” the company wrote. “Overview of Tesla’s responses revealed extra crashes that occurred in comparable environments and the place the system both didn’t detect a degraded state, and/or it didn’t current the driving force with an alert with enough time for the driving force to react. In every of those crashes, FSD additionally misplaced observe of or by no means detected a lead car in its path.”
Techcrunch occasion
San Francisco, CA
|
October 13-15, 2026
This story is growing. Examine again for updates.
