we had enough information to have prevented the sept 11th attacks. it was just buried in the noise of all the other information we'd collected. so what is our plan? create even more noise. brilliant.
Bruce Schneier likes to point out that the main limiting factor in most of these plans for gathering tons of info in the hopes of catching terrorists is the rate of false positives. eg, if you have facial recognition software in all the airports that is 99% accurate (that's better than most real software does), and you have a large airport that 10,000 people come through every day, that means that you get about 100 hits a day. that means that the security people at the airport have to pull aside 100 people a day and do a more thorough check on their id and records. since actual terrorists account for way, way less than 1% of the population, the overwhelming majority of the people that are searched turn out to be false positives. that's a pretty extreme case of crying wolf. you can imagine that after a week or two of searching 100 people a day without actually finding any terrorists, most security people would just start assuming anyone it flags is because of a software error, not because they're a terrorist.
if it were just that the TIA stuff is going to be <em>totally ineffective</em>, i wouldn't worry much about it; the government wastes tons of money on ineffective things all the time, that seems to be what government excels at. but having all that information collected and sitting in a big database opens things up for abuse by corrupt officials in a big way. it may not catch terrorists, but if you want to single out a certain person (perhaps a political enemy) and dig up dirt on them, it could prove to be very useful.
anders pearson - 2002-12-24 15:40:13
we had enough information to have prevented the sept 11th attacks. it was just buried in the noise of all the other information we'd collected. so what is our plan? create even more noise. brilliant. Bruce Schneier likes to point out that the main limiting factor in most of these plans for gathering tons of info in the hopes of catching terrorists is the rate of false positives. eg, if you have facial recognition software in all the airports that is 99% accurate (that's better than most real software does), and you have a large airport that 10,000 people come through every day, that means that you get about 100 hits a day. that means that the security people at the airport have to pull aside 100 people a day and do a more thorough check on their id and records. since actual terrorists account for way, way less than 1% of the population, the overwhelming majority of the people that are searched turn out to be false positives. that's a pretty extreme case of crying wolf. you can imagine that after a week or two of searching 100 people a day without actually finding any terrorists, most security people would just start assuming anyone it flags is because of a software error, not because they're a terrorist. if it were just that the TIA stuff is going to be <em>totally ineffective</em>, i wouldn't worry much about it; the government wastes tons of money on ineffective things all the time, that seems to be what government excels at. but having all that information collected and sitting in a big database opens things up for abuse by corrupt officials in a big way. it may not catch terrorists, but if you want to single out a certain person (perhaps a political enemy) and dig up dirt on them, it could prove to be very useful.