July 10, 2012

When I warn against Google's becoming like a god, I do not imply it alone aspires to omnibenevolence, omniscience and omnipotence. In fact, when facing inquiries on its dominance, Google itself is prompt to say strong competition disproves any accusation of omnipotence.

On cue Jason Pontin focuses the latest issue of his Technology Review on the Facebook issue. Taking Mark Zuckerberg to his word and Facebook as "a nation", he finds "we have no theory of how a virtual commonwealth should be governed, or what obligations the crown has to its people" (*).

This unfortunately is misleading. Facebook is not a nation but a non-state world actor. As such, if Mark Zuckerberg is its king, he has no people.

What Mark Zuckerberg has however, is a "Data Science Team" which "appl[ies] math, programming skills, and social science to mine our data for insights" (**). When Tom Simonite interviewed its leader, Cameron Marlow, the latter did claim "to support the wellbeing of the people who provide Facebook with their data" but let us not confuse omnibenevolence with the self-interested care of a farmer for his cattle.

While it is only natural that Cameron Marlow is there to "advance Facebook's business", Michael Wolff argues that he will fail. "The value of digital ads decreases every quarter, a consequence of their simultaneous ineffectiveness and efficiency" (***).

Yet Google's auction rules has shown how the right restraint on volume can ratchet up ad prices while harnessing the efficiency of their distribution. And while other voices do question the current effectiveness of personalized ad targeting, this is only the very beginning of its experience curve. Were we tempted to underestimate the power of Big Data, Leon Neyfakh looks at music to offer us a sobering assessment (****).

As "J. Stephen Downie, a professor at the University of Illinois" declares, "no musicologist could ever listen to 20,000 hours of music". Though individual errors are bound to occur when applying statistics and "more traditional colleagues disapprove of [digital musicologists]' replacing careful, sensitive listening with statistics", the real point is that ad targeting does not need perfection to be successful. Statistical accuracy will do.

"Think of music as a mirror", writes Leon Neyfakh. It does not reflect individual personalities only, but our very society in which Jacques Ellul (1) contended "technology should be seen as [...] an overwhelming force that has already escaped our control" (*****). Doug Hill's essay is timely. Cameron Marlow cannot claim to be "advanc[ing] social science at large", boast he can "run experiments that millions of users are exposed to" and demur "our goal is not to change the pattern of communication in society".

Why? Science is a source of truth because it delivers verifiable, factual predictions and, no sooner Man gains the means to shape the experiment, the scientist opens the way to the engineer. The latter will pick the parameters of the experiment to optimize the result predicted by the former. By its "planetary experiment", to remix Jason Pontin's well chosen title, Facebook puts Man, individually and in society, at risk of "manipulat[ion]".

Jacques Ellul wanted us to "recognize as clearly as possible the character and temptations of technique and resist them". In the instance, Zeynep Tufekci gives us the clue by suggesting Facebook is "a great resource for studying the human animal" (******). This is precisely how Facebook sees us. This is also how medical sciences study us and rightly so. But Man should not be reduced either to its animality or to any amount of pure data. Medical research draws the line by requiring each patient's free informed consent. Eprivacy is what would force Facebook to do the same.

And so Cory Doctorow echoes Ellul's call, stating "right now, the users and the analytics people are in a shooting war, but only the analytics people are armed" (*******). He calls this "a business opportunity for a company that wants to supply arms to the rebels instead of the empire".

Being behind such a company (2), I should approve but little progress will ever be made unless we understand the three challenges of this rebellion.

Downplaying the need for "ascribing property-like rights to easily copied information" is a mistake. "Nothing is ever solved by [this approach], Cory Doctorow asserts. This ignores warfare asymmetry. If, as a group, individual pirates have been little deterred by anti-piracy laws, corporate pirates are easily identifiable and hence vulnerable. Their real defense is to buy the lawmakers outright to make sure such property laws are never passed.

Cory Doctorow however correcly spotted the second challenge, auditing. Suppose corporate pirates promise to amend their way. "We can't be sure it works unless auditors descend on IT giants' data centers to ensure they aren't cheating". Even this will not work, as the FTC will discover.

And yet there does exist a "built-in compliance mechanism", the ultimate weapon for enforcing eprivacy. I have invented and patented it (3). But let us not forget the third challenge, a viable business model. While corporate pirates pay for pronaocratic protection , who is going to arm individual rebels? Until they are united and on the way to victory, they will receive no money from financial backers and no recommendation from pundits.

"Personal cofounder Shane Green" did get funded but who can blame Jessica Leber when she asks the hard questions on this start-up, which prompts its users to "upload information of all sorts [...] in a vault" upon the promise to manage their personal profile to their own benefit (********)? Yet, long term, Shane Green's vision is the only real source of sustainable growth of our Information Age.

Alas, the auditing issue remains. Bankers were once reputable people and Personal is but a data bank. What if Facebook buys it a few years from now? As long as personal data is made accessible to another party, this creates "an overwhelming force". Who believes Cameron Marlow is less honorable than Sharon Green or myself for that matter? But he has access to 800 million profiles, she to a smaller set and I to none by design (4).

My ideas cannot answer all of Jacques Ellul's fears. It certainly would protect Man from the scientific nightmare unleashed by Facebook. But I still offer a mechanical enhancement to mimick one's private dialog with one's conscience and privacy does not guarantee one's conscience will be both well formed and followed. If to reach democratic decisions it is good for votes to be secret, vote secrecy is not enough for decisions to be good.

Facebook lets science escape the bounds of reason. If I succeed by some miracle, Man will not escape the bounds of personal responsibility.

Philippe Coueignoux

• (*) ............... The Planetary Experiment, by Jason Pontin (Technology Review) - July/August, 2012
• (**) ............. What Facebook Knows, by Tom Simonite (Technology Review) - July/August, 2012
• (***) ........... The Facebook Fallacy, by Michael Wolff (Technology Review) - July/August, 2012
• (****) ......... Music by the numbers, by Leon Neyfakh (Boston Globe) - July 8, 2012
• (*****) ....... Not under our control, by Doug Hill (Boston Globe) - July 8, 2012
• (******) ..... Data Dystopia, by Zeynep Tufekci (Technology Review) - July/August, 2012
• (*******) ... The Curious Case of Internet Privacy, by Cory Doctorow (Technology Review) - July/August, 2012
• (********) . A Dollar for Your Data, by Jessica Leber (Technology Review) - July/August, 2012
• (1) for more details, see Jacques Ellul in the wikipedia
• (2) for more details, see ePrio
• (3) see my US patent portfolio, US Patents Number 6,092,197 and 7,945,954 and US Patent Application 2009/0076914.
• (4) to enable you to negociate on a personal basis, the ePrio technology does not need to give access to your profile to anyone, least of all ePrio.