Slashdot has instigated an interesting conversation about an MIT experiment that uses social data — i.e., information gleaned from social networks — to predict which students are gay. The commentary is worth a read. But the original post on Slashdot points to a Boston Globe article which summarizes the findings:

Using data from the social network Facebook, they made a striking discovery: just by looking at a person’s online friends, they could predict whether the person was gay. They did this with a software program that looked at the gender and sexuality of a person’s friends and, using statistical analysis, made a prediction. The two students had no way of checking all of their predictions, but based on their own knowledge outside the Facebook world, their computer program appeared quite accurate for men, they said. People may be effectively “outing” themselves just by the virtual company they keep.

“When they first did it, it was absolutely striking – we said, ‘Oh my God – you can actually put some computation behind that,’ ” said Hal Abelson, a computer science professor at MIT who co-taught the course. “That pulls the rug out from a whole policy and technology perspective that the point is to give you control over your information – because you don’t have control over your information.”

As the article concludes, there are a number of public policy issues worth pondering here — and ponder them we will. But there are other ways that a technology company using such data can misstep: issuing “false positives.” About a dozen years ago, at a time when I was employed as a theater producer in the San Francisco Bay Area, I became a big customer of dramatic literature on Amazon. Soon after my first order, I began getting recommendations for books on topics like, oh, gay life in San Francisco. I was not in the least offended, but I was surprised to see that the mighty and famous Amazon recommendation engine was actually quite crude (at least back then). When you automate things like gaydar — or any idea for identifying people based on their explicit or implicit behavior — you not only expose yourself to public policy concerns but also run the risk of looking like a stooge. My Amazon experience was the e-commerce equivalent of the Turing Test — the idea that artificial intelligence will reach a milestone when a machine can “pass” for a human being. The false positive in 1997 simply reminded me that Amazon customer service was driven by machines, not human beings. Not a good thing for your customers to feel when you are attempting to persuade them to stop buying from people. Of course, Amazon has spent the past ten years perfecting its engine. But not without great care and expense.