Yesterday, was released by the Committee on Technical and Privacy Dimensions of Information for Terrorism Prevention and Other National Goals (a.k.a.: CTPDITPONG). The committee was formed by the at the behest of the Department of Homeland Security to study the effectiveness of data mining techniques currently in place.
If you read this blog, you know about data mining. You also probably know that if severe distrust and loathing of data mining could be measured in people, the ACLU would be China. We've been concerned about data mining techniques since, well, forever. Remember Pentagon's failed Total Information Awareness program?
The idea behind data mining is that if law enforcement and government agencies gather enough information about every American they will somehow be able to predict future behavior and stop terrorist attacks before they happen. Now, put aside for a second how disturbing it is that our government is hell-bent on collecting as much of your information as possible. Let's focus on the fact that our national security policy is largely based on the same kind of techniques and logic that marketing firms use to figure out what kind of cereal you might buy. Sleep tight tonight, guys.
Turns out the committee feels pretty much the same way we do. Its report outlines several recommendations and, frankly, they make a lot of sense. To quote our release:
The committee made several recommendations in the report including greater external oversight of information gathering programs, a framework for both classified and unclassified programs and an emphasis on the quality, not quantity, of data.The report also discourages using behavioral patterns as a predictive measure, and considers any program attempting to assess an individual's state of mind as suspect.The committee briefed both the DHS and the National Science Foundation (the sponsors of the report) but has not received any feedback.
Ah, to have a DHS-funded report echoing our concerns so closely is incredibly validating. If only they would have listened to us years ago…
Predictive law enforcement doesn't work. It just doesn't. Even if the FBI, CIA, NSA or your local police department had all the information in the world, no one can predict behavior. Spending our counterterrorism money and resources on data mining is a waste. One of my colleagues put it another way – if these kind of predictive systems worked, data mining contractors would be making a killing in Vegas.
This report should be the definitive nail in the coffin of an ill-advised and useless policy. Keep those fingers crossed.