1

An AP article describes some recent crashes that may have been linked to the Tesla autopilot system. The article states that there have been three fatal crashes since 2016 potentially linked to the autopilot and implies it may be dangerous.

The article also states that

The company contends that Teslas with Autopilot are safer than vehicles without it, but cautions that the system does not prevent all crashes.

It Tesla right that the autopilot is safer than vehicles without it?

  • 2
    Related question: Is the self-driving car technology of 2017 less lethal than human drivers.. Note the rather extensive discussion there on the definition of terms (incidents, incidents with injury, incidents with deaths, whether autopilot was engaged at that point or not, whether the autopilot was at fault in the first place, and mixed traffic vs. hypothetical all-auto traffic). – DevSolar Jan 6 at 16:27
5

There's dispute in the numbers. Tesla makes the claim, of course, that their vehicles equipped with Autopilot have lower accident rates than their vehicles without. In 2017, the National Highway Traffic Safety Administration said that Tesla collisions dropped almost 40% after Autosteer was installed, but that number was disputed in 2019 when a report analyzing the same data said the crash rate (measured by airbag deployments) had gone up.

| improve this answer | |
  • 5
    The company that did this study countering Tesla, Quality Control Systems Corp, is a... difficult case. Their study that showed the number going "up" cherrypicked 5k cases or so from a 50k dataset, "because missing data", and made some data massaging to make it fit into the claim. More so, I'm not sure if RISKS qualify as a "reputable" periodical, so there is that. – T. Sar Jan 6 at 11:17
  • 4
    In special, QCSC is notorious for pointing fingers and saying "your data is wrong" to almost anything related to traffic data. I have no idea about their other articles, but the Tesla one at least is very weird. Take whatever they say with a grain of salt. – T. Sar Jan 6 at 11:25
  • 3
    @T.Sar-ReinstateMonica, not that Tesla isn't beyond creatively describing reality either, but in this case I'd tend to go with their numbers rather than QCSC. – Keith Morrison Jan 6 at 16:38
  • I think it's also important to look at the types of accidents that have transpired. The one that springs to mind for me is one where a Chinese driver took a nap while allowing autopilot to drive his car. It crashed into a white tractor trailer. The last time I saw an article on it, the predominant theory was that the car mistook the trailer for a cloud. Obviously the autopilot will make mistakes that humans will not, though data indicates that they make overall fewer mistakes than humans. – NegativeFriction Jan 6 at 18:07
  • 3
    The danger of self-driving cars is that it will cause drivers to think they don't have to pay attention. The accidents involving Teslas mostly share a common theme: there's good evidence the drivers weren't paying attention. There was a recommendation to the companies working on self-driving vehicles to consider systems that would detect a driver not paying attention and alert them. Tesla was the only manufacturer not to respond, which is understandable since Musk's entire marketing is to sell the idea you wouldn't have to. – Keith Morrison Jan 6 at 18:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .