You don't have to be a believer in wacko conspiracy theories to suspect that Big Brother really is watching you. From the NSA to Walmart and the Boy Scouts of America, the illusive all-powerful combine we call THEY has taken an unhealthy interest in our daily habits. Where we go, what we eat, who we see, what we do, what we buy--massive amounts of this kind of information is collected electronically every day and entered into gigantic data warehouses where it can be mined and analyzed to predict almost anything, from whether we are the kind of person who might buy Whitney's latest comeback CD to a person likely to sign up for a terrorist training camp in Pakistan.

More and more, our lives have become a series of binary numbers protected only by the limitations of the software (which is getting smarter all the time) and the human beings (who don't appear to be) that try to make sense of it. The dimensions to which all this intrudes on our daily socialization is unclear but my sense is that Donald Rumsfeld was right, at least about one thing: we really don't know what we don't know.

Not all of this is inherently evil, of course. Since 9/11 many Americans, I suspect, have become more tolerant of government intrusion into their privacy. We want our government to find and stop terrorists before they can act and so we overlook the fact that in order to do that our intelligence agents might have to listen to our conversations, too. Call your cousin in Pakistan four times in a week and start looking for the white delivery vans parked out front. But, my suspicion is that the government is now going well beyond the legal limits and that, inevitably, since there are human beings involved, serious mistakes will be made that cause rethinking of the program.

Although corporations are bound by fairly stringent rules about what information about individuals they can and cannot collect, they are circumspect about their CRM and data mining activities for both competitive reasons and, frankly, because it would probably scare the hell out of a lot of customers to be told exactly how much information the company has about them and what they might do with it.

A case study in how to do it wrong is Facebook which announced on Friday that it settled the class action challenging its now infamous "Beacon" advertising program. You remember Beacon. A couple of years ago Mark Zuckerberg, or somebody whose opinion he trusted, came up with a bright idea: Facebook would track its users' online purchases, share that information with its partner retailers, and post the information to their friends who might just buy the same thing. It didn't seem to occur to anyone that maybe some users didn't want the fact that they had bought or rented the video Debbie Does Dallas to appear in their "news feed.

In one of the more egregious examples, a man's wife found out that he sent flowers to a girlfriend. This is what happens when decisions are made by monastic geeks who are too young to remember the Robert Bork and Clarence Thomas hearings.

Facebook quickly backed down but not before several class action lawsuits were filed--one of which is the one that was settled on Friday. As part of a $9.5 million settlement, Facebook is setting up a privacy foundation, funded with whatever is left over after the lawyers get their cut (which could be up to one-third of the settlement). According to the settlement, the purpose of the foundation is to educate "users, regulators, and enterprises regarding critical issues relating to protection of identity and personal information online through user control, and to protect users from online threats."

A wrist slap, this time, but you can bet this is not the end of privacy problems for the big social networks. Two MIT graduate students opened up a whole new can of worms last week when they announced that they could identify who is gay on Facebook simply by looking at their online "friends" and using a software program that looks at the gender and sexuality of a person's friends and, using statistical analysis, makes a prediction.

The study doesn't seem to have been all that scientifically rigorous but it is troubling and raises ethical and legal questions about the big push to mine social networks for "unstructured data" which is another way of saying information that you've provided about yourself without knowing you've done so. This goes beyond the mere connecting of dots--if you bought this you will like this--and gets into a world of "inference' and undefined legal and ethical territory.

Winston Smith, Room 101, stat.

From The Blog Ate My Wiki