The author writes that, even though abuses are possible, they aren't happening and won't happen because, not only is the intelligence apparatus not interested in the mission-irrelevant data which could be abused, but those in a position to abuse are policed internally to ensure abuse doesn't happen.

Assuming this is all true, here's a thought experiment: in the event that this was no longer true and abuses were occurring, how would the public ever know? Since this entire process is impenetrably opaque to anyone on the outside, the public has only the assurances of those folks running these programs.