Originally Posted by
AllieKat
I'd absolutely agree it's less than .01% of people, yes
The thing is with big data, you literally collect EVERYTHING about a person's life and use it to create a picture. Target are masters of this, very accurately predicting things like when a woman is pregnant (admittedly, a much more common occurrence - therefore easier to create a profile of).
It's not only a much more common occurrence, but (in most cases) it's not something people are trying desperately to hide from Target or the gov't, so (a) a profile is easier to establish and (b) the women "found out" are not quite as secretive. Also, Target is significantly more competent at statistics than the TSA.
Originally Posted by
AllieKat
If you mine all the data in everyone's life, you may WELL be able to find patterns that indicate a raised risk of terrorist activity - even in bizarre things such as particular eating habits (which may come from, say, suggestions on a terrorist message board, thus explaining them). I don't doubt for one second that large scale data mining can work.
I don't think there's enough data to establish a profile. I don't think there are enough terrorists among the people being profiled (those who can apply for PreCheck) to get the numbers.
Old-fashioned intelligence work - starting from people or groups known to be associated with terrorism and following the links - is far more effective that mining all the data in every airline passenger's live looking for the needle in a haystack.
Originally Posted by
AllieKat
I seriously question whether the complete loss of privacy is worth it for a marginal improvement in safety.
The loss of privacy is so enormous and the improvement in safety so small, there's no question.