I already miss those halcyon days when if I saw a photo or video I could be reasonably sure whether or not it was computer generated.

Provenance is going to matter again, and chains/webs of trust ate always a heavy lift.

…with red light cameras.

Correct. It is a different problem.

That is actually a legitimate concern. Add it to the long list of “technologies that are cool and good except when capitalism”

Civic religion, but make it gnostic.

The SA article mentions Houston, a notoriously horrible city to drive in (concentric freeway rings, I swear). It pulled data from Texas from before 2010, and I know at least some municipalities were shortening their yellows then to encourage short stopping. It also studies changes in behavior after the punishment tool has been removed, so the drivers are already operating with years of conditioning. They even referenced the study that showed that the number of people running red lights in Virginia dropped 67% after red light cameras were installed? I need to look a little bit closer at their statistical analysis to see what confidence threshold they were using to determine that the reduction in accidents that they did see was not statistically significant, but overall I’m not that impressed by one study from a borderline pop science magazine.

Meta-analysis is the way to go.

First, don’t call it anarchy. But second, the other way to stop people from running reds is more cops.

None of these are actual problems with red light cameras, and actually people run red lights all the damn time.

There are other, newer cameras like those from Flock that run and check continuously. I prefer the old-school ones you’re talking about.