A few folks sent over this report of efforts by some researchers at UCLA to create an algorithm that can accurately take data on existing gang-related crimes, and use it to predict what gangs were involved in new crimes. It certainly has that “minority report — pre-crime” feel to it, though I can certainly see where it could be useful. What concerns me, though, is that systems like this are only as accurate as the data they use. And, as has been reported elsewhere, one of the unintended consequences of such computer analysis of crime data is that it drives police departments to falsify or change crime reports in order to make their own numbers look better. So it makes you wonder how accurate those reports will be if the incentives to fudge the actual crime data continue to be in place.

Permalink | Comments | Email This Story