Sunday, March 19, 2006

Why Data Mining Won't Stop Terror

Bruce Schneier of Counterpane wrote in Wired Why Data Mining Won't Stop Terror. "We're not trading privacy for security; we're giving up privacy and getting no security in return." Read the article you'll understand the numbers. The issue is, the system must be impossibly accurate or it will drown in a sea of false positives which still need to be tracked down and prevent agents from doing things that might actually improve security.

1 comment:

Anonymous said...

I agree with the article and the case it makes....as well as for another reason. Mining for credit card fraud....you have 100% of the data points.

Data mining for terror....you don't have all the data points. Yes I suppose you could have phone call data...and airline flight history. But it is not realistic to think you can compose a SQL statement like: select * from people where (called.Osama == true ) && (tookflyinglessons == true) || (purchasedUranium == true)

Even Isaac Asimov's science fiction version of datamining to predict the future from his Foundation Series...called psychohistory...could only do so on a macro scale.

Law enforcement probably needs better and more connected information systems....but data mining is probably not the right application.