Autopilot is a death trap. Tesla hid the bodies.
The claim you picked Autopilot is a death trap. Tesla hid the bodies.
There have been fatal Autopilot-involved crashes. Per distance driven, Tesla’s published safety stats beat the U.S. average. Both sentences belong on the card.
What is true
The name oversells. A supervised system plus a marketing department is a bad mix. NHTSA has opened investigations. That is not “the cars hunt pedestrians.”
Tesla’s Vehicle Safety Report claims fewer crashes per km with Autopilot engaged than with it off, and fewer than the U.S. average. Those figures are company-published, with selection effects (who engages it, on which roads). Independent researchers have poked holes in the comparison. Treat the exact multiple with caution; treat “much worse than human drivers per km” as unproven.
Documented failures: drivers not supervising, system confusion at crossing traffic, parked emergency vehicles. Some crashes are ghastly. The ethical issue is the gap between the word “Autopilot” and the legal requirement to pay attention. You can want stricter naming and driver monitoring without claiming a unique slaughter. Compare to the 40,000 U.S. road deaths a year, almost all of them not in a Tesla.
Key figures
- Company stats
- Fewer crashes per km with Autopilot than U.S. average
- Caveat
- Self-reported, road-mix bias, naming oversell
- Real issue
- Supervised system, fatal misuse, regulator interest
