I should clarify my opinion encompasses both red light and speed cameras.
I know when traffic cameras are used never 100% of violations caisght on camera are used. The jurisdiction has to experiment on the % of violations enforced. Too many and the public gets angry and votes you out of office. Too little and it is not a deterrent. So in some cases, these municipalities where they have been deemed not effective have neglected to do this optimization and started with a % too low of enforcements.
People just assume traffic cameras are a cash grab. But the purpose is to deter bad driving behaviour at a lower cost than having officers in the field. Fines are the stick, but if it means money for municipities to reinvest that into the city, then great. For each city, there will be a threshold of % violations enforced before it be comes effective.
You're right - there's data to show cameras can make some intersections safer and make other intersections less safe. It isn't simple. There's also some pretty clear cases where the cameras were strictly a cash grab. Cities have been caught shortening the yellow light duration when they don't meet their revenue numbers.
So I don't think you can start with the assumption that traffic cameras make us safer. If they don't necessarily make us safer AND the criminal process is unconstitutional, it gets easier to oppose the way they are used.
I won't argue that traffic cameras are good news from a revenue perspective, but I'm not sure it's effective. Check out:
https://www.motorists.org/issues/red-light-cameras/studies/.
The references on that page make the argument that cameras increase the number of accidents.