] Police in Florida are creating a counterterrorism
] database designed to give law enforcement agencies around
] the country a powerful new tool to analyze billions of
] records about both criminals and ordinary Americans.

Mini-TIA?

The problem with systems like this is not so much the information they have but the sort of questions you are allowed to ask them.

Asking who has brown hair and a red truck within a 20 mile radius, in the context where this is a suspect description in a murder, is a standard question that police ask all the time. Having this information more readily available is probably a good thing (unless you're an anarchist).

However, if you run a correlation which shows that people who have brown hair and red trucks are 30 percent more likely to commit murders then average, and subsiquently decide to submit people fitting that profile to additional scrutiny at airport security, you've crossed into pre-crime, and that is where the policy debate lies.

This question is going to continue to be raised. Poindexter, for all his faults, is a leader. He is way ahead of the curve. We'll see a lot more of this over the next 20 years from all kinds of directions, just as we are seeing similar techniques used in unrelated fields (Customer Relationship Management).

Objective research into the effectiveness of pre-crime, and the impact of it upon innocents, is sorely needed. Unfortunately, finding objective researchers is going to be damn near impossible. On the one side we've got arms dealers and drug smugglers, and on the other side we've got civil liberties advocates. Much like the studies on RF related cancer, the truth is probably only going to be found somewhere in the dialog between them, in an environment where both sides are given the resources they need to do the studies they want to do, and there is absolutely no political pressure to rush things into application.