The problem is that the scale messages makes even seemingly small false positive rates a big problem. Suppose you had a false positives only 0.01% of the time

The problem is that the scale messages makes even seemingly small false positive rates a big problem. Suppose you had a false positives only 0.01% of the time -- a ludicrously generous estimate. When you are looking at hundreds of millions or even billions of messages per day, that is far too much noise for the system to be useful. Maybe the FBI/NSA is so good that they never see false positives. I doubt it, though, given how many non-terrorists have been harassed at airports, had their peaceful protest groups infiltrated by undercover agents, etc.