August 14, 2007

Food, energy, data. A radical increase in their availability launched our agricultural, industrial and information ages. Since overabundance of food begets morbid obesity and cheap energy encourages pollution, one might expect people to be wiser to the downside of too much data. Alas, when has hindsight been the basis for foresight? Among the incessant paeans to King Data, the unquenchable quest for more data, it is quite comforting to read some sober accounts which dare to question data doters.

Christopher Caldwell reminds us that most of the information posted on the Internet is bad data, from teenagers who strut online to CEO's who self promote under pen names (*). Inspired by the old adage "in vino veritas" to embrace a second life of bartending, Michael Swaine makes fun of "the notion that the cure for bad information is more information" (**). His recommendation? Metadata. Whether bad data is filtered out by accreditation, credentialing, metainformation or a recommendation mechanism, never mind the bottle and enjoy the liquor(1). Too bad good recommendation systems are so hard to build and maintain. Apparently what I called the seventh note in my Pythagorean hymn to privacy is a very high note to sing.

Bad data is bad because refining it into truth wastes our time, which is finite. Good data too has the power to hurt us. When it is our personal data, good data is the target of everyone else while our right to privacy is limited. Wits will wisper, let us simply turn our personal data into bad data. Are teenagers and CEO's up to something? Must we elbow our way into the overbooked flight from responsibility? To tackle the problem, why not start with Keith Bradsher's report on China's experiments with smart residency cards in Shenzhen(***). "Data on the chip will include [...] work history, educational background, religion, ethnicity, police record, medical insurance status and landlord's phone number". And this is just the beginning.

Now we can recoil in horror and condemn citizen carding as the work of the devil. But think again. First the US can hardly give China lessons, especially in matter of surveillance (07/03/07, 08/07/07 fillips). Second the facts listed by Bradsher, which amount to a good personal profile, merely reflect what is already known about us. Whether my profile is on some central servers out there or a personal card in my hand, it has already been compiled.

With good personal data, it is not so much the "where" which ought to be the issue as the "who" and the "how". Who has accessed to my profile and what control do I have over it? Wouldn't one good card in hand I really control be far better than a few dozen servers so remote I do not even know they exist? What if the police could read nothing more from my card than my name unless I agree to it in the presence of my lawyer? What if I could insert it in my computer and let marketers target my confidential profile without ever giving them access to it? What if marketers were forbidden to ask third parties for any fraction of my profile without my explicit consent?

Conveniently putting our profiles on smart cards would expose our present absence of data rights, no matter our country. It may even shame some countries into giving them back to us.

Philippe Coueignoux

• (*) ..... Not Being There, by Christopher Caldwell (New-York Times Magazine) - August 12, 2007
• (**) ... Hair of the Bit that Bit You, by Michael Swaine (Dr. Dobb's Journal) - September, 2007
• (***) . China Enacting High-Tech Plan To Track People, by Keith Bradsher (New-York Times) - August 12, 2007
• (1) "Qu'importe le flacon pourvu qu'on ait l'ivresse", Alfred de Musset