While the threat (expressed as P(harm)*harm) from XYAgency surveillance is low or medium, the unmitigated threat from the things their surveillance practices pr

While the threat (expressed as P(harm)*harm) from XYAgency surveillance is low or medium, the unmitigated threat from the things their surveillance practices protect against - namely, terrorism - is also low. The mitigation effect is lower still. Objectively, even in 2001 terrorism was a negligible risk compared to everyday risks, and subjectively, there hasn't been a major terrorist attack for years. The only reason mass surveillance exists is a rationale by the US state that "more power is good". This may apply to the US army, but not to the spies. Also, like you say, once the american spies start to (pun intended) "Interfere in american elections", then the problem affects everybody with P=1. Personally I could live with it trump were the only candidate they work against, but if they do it to trump, they likely do it to others. (I see no evidence of interference or other abuse of collected data currently, but I think it's dangerous to give them the power to collect all this info that can in principle be abused for selective prosecution and/or blackmail). Also, it affects not just politicians, but also corporate execs, who can be further pressured using the other means of the state, and who themselves have power the state can deputize. A democracy should have the surveillance powers that are proportionate to the benefit from these powers and no more. There is a positive value in minimizing state surveillance power; this concept seems lost on america. (In fairness, it seems lost on conservative parties worldwide.)