On October 2, 2023, a human-driven car struck a pedestrian in San Francisco and threw her into the path of a Cruise vehicle. My first post focused on whether the robotaxi could have avoided that initial collision. That missed what later became the central safety issue.
An NHTSA consent order says the Cruise vehicle stopped, then resumed movement and dragged the pedestrian about 20 feet. NHTSA also said Cruise’s first required reports omitted that post-crash behavior. California’s DMV suspended Cruise’s driverless permits on October 24, 2023.
I have ridden Waymo in San Francisco, and I still think driverless service can be useful. This case made me look harder at what the vehicle does after an impact and what the operator tells regulators afterward. A car can be involved in a crash it did not initiate and still make the injury worse. The reporting failure then makes it harder for anyone outside the company to judge the system.
The operational test for a fleet is not only whether it detects obstacles before a crash. It is how it stops, calls for help, preserves evidence, and reports the full sequence when things go wrong. That is why I separate actual robotaxi service from launch language.