Facebook espionage

This was an interesting NY Times article about how the U.S. government is looking to expand it's intelligence and counterterrorism groups into social networking:

As social media play increasingly large roles in fomenting unrest in countries like Egypt and Iran, the military wants systems to be able to detect and track the spread of ideas both quickly and on a broad scale. The Defense Advanced Research Projects Agency is soliciting innovative proposals to help build what would be, at its most basic level, an Internet meme tracker.

It would be useful to know, for instance, whether signs of widespread rebellion were authentic or whether they were being created by a fringe group with little real support. Among the tools the successful seeker of government funding might choose to employ: linguistic cues, patterns of information flow, topic trend analysis, sentiment detection and opinion mining.

It's an ambitious goal, but I wonder if this type of thing is more art than science. If someone could accurately predict future trends with a $42 million piece of software (the amount the U.S. government is putting up to sponsor research in this area), then I'm sure that someone would be putting the software to a more lucrative, business or stock related use. The only people who seem likely to volunteer to do such an ill-fated project are the ones who either (1) know that they'll never come up with the software anyway, but don't mind getting paid to spin their wheels or (2) double agents who know that they'll come up with the software and will sell the U.S. government out. But that's the problem with big government is that the bigger it gets, the harder it is to tell what's going on -- right libertarians?

A better use of funds might be to hire young sociopaths to monitor social networking sites and push people into doing one thing or the other. Of course, they could always go "bad" and engage in behavior counter to official policy. Easy solution! Just hire sociopaths to monitor those ones, and so on.