Tesla has started to release its quarter Autopilot safety reports after stopping for a year. The automaker claims some improvements.
Since 2018, Tesla has been trying to create a benchmark for its improvement in Autopilot safety by releasing a quarterly report that compares the number of miles per accident on Autopilot versus off Autopilot.
The data was always limited and criticized for not taking into account that accidents are more common on city roads and undivided roads than on the highways, where Autopilot is mostly being used.
However, it was still helpful to compare it against itself over time and see if there were any improvements, and there were some incremental improvements at times.
Tesla suddenly stopped releasing those quarterly reports in 2022 without any explanation.
Now, the automaker has decided to start it up again and released reports for all quarters up to Q3 2022.
Tesla wrote for the most recent data:
In the 3rd quarter, we recorded one crash for every 6.26 million miles driven in which drivers were using Autopilot technology. For drivers who were not using Autopilot technology, we recorded one crash for every 1.71 million miles driven. By comparison, the most recent data available from NHTSA and FHWA (from 2021) shows that in the United States there was an automobile crash approximately every 652,000 miles.
That’s compared to one crash for every 4.35 million miles driven with Autopilot technology back in Q4 2021 – the last time Tesla was regularly releasing the data.
Tesla was also kind enough to plot the data into a chart this time for better visualization:
As you can see, it goes up and down, but that’s partly seasonal. There are notoriously more accidents during the winter due to road conditions and because it gets dark sooner.
Since Tesla stopped gathering the data, the automaker has also significantly grown its Full Self-Driving Beta program, which actually allows the use of more Autopilot technology on city roads.
However, it’s unclear whether Tesla includes this data in this report.
Top comment by brucet
Fred, do you know if there are data available of the frequency of accidents among all cars of similar age to Tesla Autopilot cars (2014 or newer) occuring on highways?
That comparison to accident frequency for all vehicles in the US, including old and poorly maintained ones, and accidents that occur on city streets and secondary roads where Autopilot can't be used, always struck me as unfair.
I know. It’s a very limited dataset, and I too wish Tesla would be more transparent. But it’s the best we have right now, and it does show some improvements.
That’s what we have to work with for now.
As I recently reported, I genuinely hope Tesla would release more data specifically about its FSD Beta program so we can start seeing some solid numbers building a path to Tesla delivering on its self-driving promise.
The company restarting to release these reports might be a step in that direction. We will see.
FTC: We use income earning auto affiliate links. More.