> That geofence warrants scoop up innocents’ data has now been proven beyond doubt.

While true, it's unclear how relevant this is. The equivalent police process of canvassing people determined to be likely bystanders also involves police interviewing and investigating people who turn out to be innocent.

Probably the relevant question is "Is there a fundamental difference between police identifying potential involved parties via eyewitness testimony and 'common sense' practices like looking for windows overseeing a crime scene and police fetching location identifiers from passive sensors?"