Autonomous Vehicles Built to Report Teenagers to Police

Two teenagers in a Waymo driverless taxi in San Francisco were drinking and firing toy guns when the vehicle's systems detected the disturbance, disabled itself. Sent police to their location.

The incident became a privacy story almost immediately. Critics asked how a corporation could hand teenagers over to law enforcement, how much surveillance happens inside these cars. Who decided that behavioral intervention meant cop involvement.

The entire debate rests on an assumption nobody's actually defending. That assumption is that Waymo chose to call the police. That choice language matters because it implies discretion. That Waymo's safety team evaluated the situation, made a judgment call, and decided this particular incident warranted law enforcement.

But that's probably not what happened. Insurance carriers don't work that way, and neither do regulatory approval bodies. Autonomous vehicles operate under conditions set by underwriters and government agencies that are terrified of liability. Conditions that almost certainly mandate automated escalation protocols. When a vehicle detects certain behaviors, the system doesn't pause for human deliberation about whether police are proportionate. Instead, the detection triggers the response.

This distinction collapses the privacy scandal into something larger and more structural. It's not a corporate privacy violation. It's evidence that the industry's foundational architecture has already made this decision for everyone. The question isn't whether Waymo should have called the cops on those teenagers. The question is whether anyone operating autonomous vehicles in this liability environment can actually not call them and stay operational. If the answer is no, then we're debating whether the entire insurance and regulatory infrastructure that makes autonomous vehicles possible has already decided civil liberties questions on our behalf.