End of Privacy

By John Parkinson  |  Posted 08-06-2007
Here's a quick Trivial Pursuit question: How many data items would it take to uniquely identify you from among the other 6.7 billion people on earth? The answer is at the heart of the debate over personally identifying information, which in turn is at the center of the evolving information privacy debate. Let's make it a little tougher by disallowing all direct personally identifying information: no names, no Social Security or driver's license or passport numbers, no fingerprints or retinal patterns or DNA profiles.

The answer might surprise you. A combination of just gender and the U.S. five-digit zip code (or its foreign equivalent) for your address would on average eliminate all but about 35,000 people. In most zip codes, a date of birth would narrow it down to around

95 people. That's just three data items, none of which would generally be regarded as unique to you, and you're down to fewer than one in 100.

If we add in some situational data—facts about you that don't identify you specifically but build a recognizable context around you—the kind of car you drive, the restaurants you frequent—I can typically identify you with just a couple more items, usually a dozen max, none of which would be considered personally identifying information. It's this ability to build context and use it as an efficient information filter that makes privacy so hard to maintain.

We all leave a trail of data items as we move through the world, and we always have. Technology has simply made it easier and cheaper to record and analyze these traces. Today, for about half the world, there is no real privacy. The key questions, therefore, become: Who owns our personally identifying information? Who assures its accuracy and relevance? Who can access and use it? What are its permitted uses? Too many of the answers depend on where you live and how the laws there constrain or allow data use. This leaves businesses and technology managers facing some complex issues even beyond the ethical debate on how the information can be used.

But as bad as things have gotten from a privacy viewpoint, they're about to get a lot worse. As the world becomes more routinely instrumented (think E911-enabled cell phones, GPS, WiFi access points and black-box recorders in autos and surveillance in the name of public safety), event correlation software will make it possible to construct a nearly complete record of your life and make it very hard to hide. This can be a blessing if you have to prove where you were (or weren't) at some point, but I'm not sure we as a society are ready for this level of transparency. And as information managers, we have to be careful, where appropriate use is not yet defined, to avoid making post hoc decisions on what can and can't be done with the data our systems collect.

For CIOs, that means staying on top of the debate and getting some appropriate policy defined—or at least getting a discussion of the issues under way internally—even if you have to modify process and practice later.

Determine what compliance and audit requirements you'll have to meet before you have to meet them. Consider training and awareness needs. And add one more item to the long list of concerns demanding your attention.

John Parkinson has been a business and IT consultant for more than two decades. Please send questions and comments to editors@cioinsight-ziffdavis.com.