The article seems to be more interested in being outraged in principle than demonstrating anything objectionable specifically. The law in question seems to requ

The article seems to be more interested in being outraged in principle than demonstrating anything objectionable specifically. The law in question seems to require cars to implement a way to detect dangerous driving and be able to respond to this (presumably by forcing the driver to stop). There is nothing suggesting a need (or permission) for data collection beyond the specific incident or for law enforcement involvement at any point. It's worth mentioning that cars with "autonomous" driving modes or intelligent assistance already record similar data so the data generation itself does not seem to be the issue here. The far more troubling question is a) how the car is supposed to determine you're driving dangerously (and how likely false positives are) and b) how the car is supposed to safely come to a stop in a way that does not endanger other people (hopefully not by just turning on the hazards after coming to a full step on the left lane of a highway). Given that I recently had a car blare at me because it wouldn't believe that I was looking at the road no matter how I turned my head, where I looked and how much I opened my eyes, the prospect of a car deciding its driver is drunk and doing something dangerous seems like a much bigger concern than law enforcement or eco terrorists hacking my car to prevent me from driving to Starbucks.