Often it's the impression, threat, or procedural presumed validity of effectiveness that matters far more than actuality.

The history of debunked methods used in government regulation, court and criminal evidence, or business practices is high. Simply posting "this area under visual surveillance" changes behaviours, and animal studies show dramatic differences in behaviour and detectable stress indicators under conditions of surveillance or predator presence. Even alpha predators react strongly negatively to the presence of human voices:

https://www.theatlantic.com/science/archive/2019/07/humans-p...

Vigilance stress is almost certainly deeply innate. Humans are just as much affected:

https://www.healthline.com/health/hypervigilance

https://www.healthcentral.com/article/hypervigilance-in-anxi...

But one of the easiest to track everyone everywhere. What makes it so dangerous is its scaleability. The large error margin isnt a problem if your goal isnt finding single terrorists but create a surveillance state.