- Tesla claims its electric cars driving on Autopilot are far safer than those driving without it.
- In the third quarter of the year, Tesla recorded one crash for every 7.08 million miles driven on Autopilot.
- By comparison, there's a vehicle crash every 670,000 miles driven in the U.S., Tesla said citing federal data.
Tesla has released its Autopilot safety report for the third quarter, indicating a significant increase in the number of miles driven using the technology before recording a crash.
Autopilot and the more advanced Full-Self Driving software are critical to CEO Elon Musk's ambitions of deploying self-driving cars and Robotaxis in the future. That's why the safety reports—if accurate—are an indication of how far ahead the company has come in improving this technology.
In the third quarter, Tesla claims to have recorded one crash for every 7.08 million miles driven using Autopilot, an advanced driver assistance system (ADAS). For drivers who were not using Autopilot, Tesla recorded one crash for every 1.29 million miles driven. Tesla collects this data from millions of EVs on the road today that use these systems.
That's a big jump compared to the third quarter of last year, when Tesla recorded a crash for every 5.8 million miles of driving on Autopilot. By comparison, Tesla said there was a crash in the U.S. every 670,000 miles driven, citing data from the National Highway Traffic Safety Administration (NHTSA) and the Federal Highway Administration (FHWA).
Autopilot is standard on all Teslas, whereas FSD is an $8,000 option. They're both Level 2 ADAS systems, meaning driver supervision is required.
Autopilot can accelerate, steer and brake on its own but requires drivers to keep their hands on the steering. FSD is a more advanced version where drivers can take their hands off the wheel, but must supervise and intervene if the car falters.
Experts told InsideEVs before the Robotaxi reveal that Tesla's safety reports are statistically insignificant because the specifics are obscure. The reports don't specify the severity of the crashes, whether they were fatal or non-fatal, whether driver or vehicle error was to blame, or the driving conditions involved. And most importantly, the reports don't include any FSD data, which is the underlying tech for the Cybercab.
By comparison, NHTSA and FHWA safety reports are far more detailed. NHTSA does not provide the exact figures on the number of miles driven before a non-fatal crash as readily as it does for fatal crashes but provides a sample size large enough to indicate how safe (or not) our roads really are. Data from NHTSA and National Safety Council indicates 1.33 deaths per 100 million miles driven. That implies humans already drive 99,999,999 miles before a fatal crash occurs.
Tesla has not provided comparable data for Autopilot or FSD. Plus, Autopilot and FSD have been linked to dozens of deaths and hundreds of crashes. Just a week after the Robotaxi reveal, NHTSA opened another investigation into 2.4 million Teslas after FSD's alleged role in four reported collisions, including a fatal crash.
Even if Tesla's quarterly safety reports are to be believed, they indicate little about how safe the systems really are for a larger audience on a national level in different driving conditions.
Have a tip? Contact the author: suvrat.kothari@insideevs.com
