I'm of the opinion that once enough people get fed up with a surveillance state, or even a surveillance society since private entities are involved, that the best way to "fix" the problem is by collectively generating noise that makes it too expensive and time consuming to find a needle in a haystack. Right now they probably generate very few false positives, however if many people went out of their way to actively generate false positives on a regular basis, you've effectively disabled such a system and manufactured reasonable doubt.
Generating deliberate false-positive inducing noise in communications deemed to be private between two or more individuals who know one another should be protected as free speech. To argue otherwise would be the equivalent of prosecuting an individual for yelling "Fire" in their own home among friends and stating that such an act is a clear and present danger to the US.
IMHO automated cooperative manufactured reasonable doubt will probably be one of the last bastions of civil liberties in a surveillance society.
If this privacy-invading-data-miners are using computers and mathematics they’ll surely find better and better ways of filtering the false positives. But before it even comes to that they’ll probably just scare people into submission by making examples of people who create false positives--probably by treating them as true positives.
there's a problem regarding false positives though. A portion of those may be guilty of other things.
There was this interesting article on this idea of terahertz lasers in airports. I think these machines are great because they are programmable and specific. You can program them to look just for explosives and this reduces the search issues significantly. But what of the fact that they would mean the TSA might be Constitutionally barred from looking for drugs? Would this retard adoption?
I think if you want to show you are doing a great job at law enforcement and minimize the warrant requirement, you want to have as many false positives as the courts will let you get away with. "Yeah they only found a few oz of pot, but they had probable cause to believe he was a sexual predator, so the evidence is admissible."
There are three problems with this I can see straight off the bat:
- The noise itself may be interpreted as incriminating. If someone wants to make trouble for you, they can, based on the noise. Yes, you have plausible deniability, but this costs time and money.
- Fuzzing signals is tricky. If someone's snooping for unspecified suspicious behavior, noise may cloud things. If they're looking for specific data to tie you to people, places, times, events, etc., there are very powerful tools to cut through the things you're not interested to just the stuff that's relevant. Methods of masking printer identification marks suffer a similar problem.
- Even if you're generating pure random white noise, under a regime compelling decryption on request, you've now got to make the case that noise is in fact noise, and not very securely encrypted data. Again at a cost of time/money in the face of someone who wants to make trouble for you.
Generating deliberate false-positive inducing noise in communications deemed to be private between two or more individuals who know one another should be protected as free speech. To argue otherwise would be the equivalent of prosecuting an individual for yelling "Fire" in their own home among friends and stating that such an act is a clear and present danger to the US.
IMHO automated cooperative manufactured reasonable doubt will probably be one of the last bastions of civil liberties in a surveillance society.