X's recent post on data mining makes me think, and I'm going to be a bit contrarian if I may. Bear with me as I digress a bit.

A friend of mine recently finished up a job, very high paying, for a Bush-donor owned company that contracted him to survey archaeological sites in Iraq. The purpose? To assess areas "best suited for the construction of luxury hotels and resorts for Western tourists." You know, see some ruins, do some shopping, enjoy your own kaffir servants for a week. It will shock you (not) but the project has been sidelined "temporarily," until the security situation in Iraq makes such investment sensible. Which will be shortly after hell has frozen over, but you knew that.

Anyway, my friend put together rather comprehensive data for the corporate project manager, fine details about ancient sites and their potential as tourist draws. The project manager was responsible for putting together the executive summary for the higher-ups, and my friend, despite being a pointy headed academic, writes clearly and lucidly. Even the barely educated can make sense of his work. So how did the executive summary read? My friend was horrified to review it, and noted that the project manager had completely mangled the data. She went so far as to insert totally ridiculous unfacts, many of which directly reflected her own ignorance of Middle Eastern and ancient history, as well as her fundamentalist christianist bias. The short version is that she ditched most of the academic work and inserted her own fantasy, the better to impress her corporate masters.

Now, if you're wondering what this has to do with data mining, I'd just like to remind everyone of a couple of facts. Fact 1: Bush's government is breaking new records for cronyism, a la Brownie. We've probably never had so many directly political, utterly unqualified, generally clueless government officials as we have now. Fact 2: The government, across the various agencies and military groups, do not truly have the "finest minds" when it comes to computers. See: the recent losses and thefts of computers at the VA, State, and Homeland Security. Fact 3: It takes a great deal of time, effort, and intellectual ability to mine data in a meaningful way. Data is useless when those who review it have already made about their minds about what it says and means. One cannot avoid empirical methodology and at the same time produce the results it guarantees.

I am very worried about bright, competent, politically motivated people in my government mining my data, building up files about my activities, and otherwise spying upon me. At the same time, I'm not convinced that those setting up all these mining programs I keep reading about are actually reviewing the material in a worthwhile fashion. Data analysis is complex, painstakingly time consuming, and requires a scientific intellect. Folks like that don't seem to be running a lot of government organizations of late. Or at least many at the executive levels are not.

I don't mean to undermine Xenophon or Lambert or any other writer who documents the atrocities with respect to privacy and government intrusion. But it seems to me I should be more worried about a group of people too stupid to understand that torture doesn't produce solid intelligence, but who rely on it as a primary method of intelligence collection anyway.

They are brutal, ignorant and ahistorical. That's what makes them truly dangerous.