Possibility and Probability

A Python programmer with a personality thinking about space exploration

16 October 2006

Protecting minors by mining MySpace

by Nick

This is a very interesting article:Wired News: MySpace Predator Caught by Code Finally, a mashup that does something useful. :)  As the article points out there are certain patterns of usage that probably would raise red flags. What surprises me the most though is people aren’t excited about this type of technology being used. Given the large size of the user base of MySpace, it is pretty impractical to have a set of eyes on every user and every posting that goes on. Granted, there is a real possibility of false hits when doing an automated search like this. But given that a computer can sift through the set of millions of possible hits and narrow it down to a few hundred (which can then be followed up by a human), to me it is a no brainer. The best part of doing an automated data mining scan of a site like MySpace would be that its a computer, not a person doing the scanning. Computers don’t make judgements, or laugh at your music choices, they just scan. To me it seems like this would be the best of all possible worlds: Let the computer make the rough pass over the site, and pick out the most “questionable” users/postings for human followup. That way the “invasion of privacy” (if there even is such a thing on the internet anymore) is limited. And by making sure that a human is doing the follow up, we hopefully remove the problem of the over-zealous filter that assumes everyone is bad. I’m curious to see what the reaction is when the reporter releases the code that lead to the investigations in the story. Will MySpace adopt its usage? Will vigilante surveillance groups pop up and patrol the internet? Interesting times lie ahead…

tags: