What I find interesting is the thought experiment "how would you act if you wanted to evade the surveillance?" Simply going "off grid" won't work b/c it provides nearly as reliable a signal to those doing the tracking.

I consider it likely that terrorist organizations have thought about this, and wonder what approaches they are using to try to thwart it.

Since it would seem fairly possible to create a convincing appearance of legitimacy via FB and Gmail and even banking, the value of the surveillance must lie in understanding human network topologies so that physical agents can be deployed strategically to do traditional humint.

Since most of the data is obtained w/o warrant it is unlikely to be admissible and used toward a conviction. This would be a problem if convicting the individual whose data was scanned was the goal. The goal is to aid the efforts of traditional law enforcement and intel by generating something like a "suspicion heatmap" over the social network as revealed by our financial, email, and social network transactions.

This helps efficiently deploy agents, but once deployed the agents still need to make friends and infiltrate their assigned network edges. Cover is not always necessary to do this, since many members of the network are upstanding citizens who are happy to help with something security related.

As I've speculated earlier, I would expect that honeypots (large and small) are the preferred way to actually perform experiments on the suspicion graph.

It's fascinating to think about what sort of honeypots probably exist.