The Transparency Trap

I’m watching the NHTSA recall feed again. It’s a constant stream of safety complaints, investigations, and remedies. Today, a software defect in a Level 2 autonomy system triggered a recall affecting 187,000 vehicles simultaneously. One line of code touches two hundred thousand cars at once. That’s not a bug fix. That’s a lever.

A tech reporter from Reuters just pinged me about it. She wants to know if this is a sign of deeper issues with autonomous vehicle safety. I told her it’s a sign of deeper issues with public perception. The same transparency that’s supposed to prove AVs are safer is what makes every defect look like proof that they’re not. You can’t win this game.

We replaced one failure mode—human error—with another: algorithmic error. Except the second one scales infinitely faster. It’s not the defect that kills you. It’s the coordination. A human driver crashes once. An AV fleet has to be fixed while it’s still driving people.

Every person not in one of these cars is voting against it every time it fails. You’re not just fixing a bug. You’re running a referendum every 60 days. We built better and better detection systems. Now we catch problems we never would have seen before. Which means AVs look worse, not better, the more honest we are about them.

This isn’t a story about whether AVs are safe. It’s a story about what happens when you make a technology transparent enough that every defect becomes political.