Support CleanTechnica’s work through a Substack subscription or on Stripe.
It’s funny how things line up in the fabric of time at times. A few days ago, I wrote an article about Tesla having perhaps 29 supervised robotaxis in operation in Austin, Texas. It’s not clear what the exact number is, and it’s not clear exactly how many crashes these robotaxis have been in, but one of our readers, Ole Laursen, directed me to an NHTSA site that does provide some accident data … sort of. I was just starting to write about this, and as part of that discuss the need for more info and better data, and then I read this piece from Steve. That does a much better job of explaining what is lacking, but I still want to dive into the stats Ole shared.
First of all, note that I’m highlighting crashes related to two different things — ADS and ADAS. Here’s how the two (ADS first and ADAS second) are explained by the NHTSA (bold emphasis added):
“Automated driving systems, still in development, encompass SAE Levels 3 through 5. In its mature state, a vehicle equipped with ADS aims to perform the entire dynamic driving task on a sustained basis within a defined operational design domain without driver involvement.
“Level 2 advanced driver assistance systems provide both speed and steering input when the driver assistance system is engaged but require the human driver to remain fully engaged in the driving task at all times.”
Additionally, the following are the NHTSA’s reporting requirements on these two different solutions:
“ADS: Entities named in the General Order must report a crash if ADS was in use at any time within 30 seconds of the crash and the crash resulted in certain property damage or injury.
“Level 2 ADAS: Entities named in the General Order must report a crash if Level 2 ADAS was in use at any time within 30 seconds of the crash and the crash involved a vulnerable road user being struck or resulted in a fatality, an air bag deployment, or any individual being transported to a hospital for medical treatment.
“Crash reporting criteria are stricter for ADS than ADAS to account for the higher level of automation in ADS and driverless operations in surface transportation. More information on the reporting requirements is available in the General Order.”
Also note that the data is through October 15, 2025. On to the data (before coming back to its limitations).
On this first one, I’m a bit surprised by how much more Tesla vehicles with Autopilot or FSD (Supervised) activated have crashed. Are drivers of the other vehicles not using their ADAS options as much and thus not crashing with them on as much? Or are they using them more carefully and wisely, whereas Tesla drivers expect too much from the technology and thus get into far more (thousands more) accidents? Other automakers generally have very clear places where ADAS is useful, whereas Tesla drivers may think they can (or even should) use their ADAS everywhere, even where it’s not really ready for us.
Naturally, as Dr. Jonathan Slotkin noted, “We need the denominator, not just the numerator.” We don’t know how many miles these vehicles were driven with ADAS activated, just how many times they crashed. One would think other automakers would log plenty of ADAS miles, since there are millions more of those vehicles on the road, but as I said, Tesla drivers are probably more likely and certainly more encouraged to use ADAS widely and often. So, we don’t know if other automaker vehicles have logged many more ADAS miles than Tesla or many fewer miles, or about the same.
It’s basically the same dilemma here. Waymo has certainly driven the most ADS miles by far. But how many ADS miles any of these companies have driven, we have no idea.
Circling back to Tesla, it appears to have only 29 robotaxis on the road in Austin, and they haven’t been there for long. Cruise and GM probably have duplicate reporting going on, since Cruise was owned by GM. Interestingly, Cruise seemed to be neck and neck with Waymo for a while, until a very unfortunate and odd accident led to the company shutting down its robotaxi operations, but Waymo has scaled up tremendously since then. What we can say with a decent amount of confidence is that Waymo has driven tremendously more ADS miles than anyone else. Are its number of crashes still high relative to miles driven? Who knows? We don’t have the denominator.
At the end of the day, we don’t know if these stats are good or bad relative to humans, because we not only lack total number of miles driven, but we also don’t know where these crashes occurred (types of road, city, etc. can have a big difference on likelihood of a crash). As Slotkin (and Steve) noted, give us more data!
Update: Just to be clear, I want to highlight the overall message here, as succinctly put by a reader after the article was published:
To be clear.Those raw numbers, without the context of the denominator, have no meaning whatsoever.
To attempt to draw any conclusion without further context is faulty logic.
Sign up for CleanTechnica’s Weekly Substack for Zach and Scott’s in-depth analyses and high level summaries, sign up for our daily newsletter, and follow us on Google News!
Advertisement
Have a tip for CleanTechnica? Want to advertise? Want to suggest a guest for our CleanTech Talk podcast? Contact us here.
Sign up for our daily newsletter for 15 new cleantech stories a day. Or sign up for our weekly one on top stories of the week if daily is too frequent.
[embedded content]
CleanTechnica uses affiliate links. See our policy here.
CleanTechnica’s Comment Policy
Share this story!