I can see how that might work, but signal reception is always noisy. I doubt RSSI would be a reliable measure. You could partially wrap the AirTag in a scrunch

I can see how that might work, but signal reception is always noisy. I doubt RSSI would be a reliable measure. You could partially wrap the AirTag in a scrunch of alfoil and it would mean every movement of the phone would massively change the reception, it would look like a variable distance. Also, I wonder if it is a fixed time with no overlap? Because you could certainly track someone, eg through a shopping centre, by seeing when a beacon turns off and then listening to new beacons. Correlation would be trivial. And if the e.g. 30m clock is accurate then you could reidentify hours later by just listening to the rollover time, so they would have to vary the rollover at least. I wonder about false alarms, because you can easily sit on commuter train for an hour and have someone next to you, even more so for long distance travel.