Computers, Privacy & the Constitution
-- WardBenson - 27 May 2009

A Return to Calvinist America

During a recent argument I had some difficulty defending the anti-data-mining cause. Ok says the 1L, putting aside fears about abuse by employees of data aggregators, the consequences of inaccurate records, and the potential loss of autonomy that may occur depending on how much you buy behavioral economics, if the data aggregators get it right, won't that be more efficient for the economy and the nation? What if the banks really can accurately determine everyone's true credit risk? What if the health insurance companies really can predict someone's lifetime health care costs? What if New York law firms can hire only those people who will work the hardest and not leave after immediately paying off their debts? Won't everything work more smoothly? While at the time I failed to give a convincing response, as is typical, a better (in my opinion) response came to me about ten seconds after the conversation ended. Sure, it is possible that through the use of data-mining and analysis banks may be able to much more efficiently dole out credit and companies will be much better able to know who they are hiring. However, no matter how much data-mining may be good for the economy (at least in the short-term), it poses a threat to one aspect of what America stands for. Crucial to American culture and mythology is the belief that America gives people a unique opportunity for reinvention. Whether their model be an immigrant who leaves behind the role he held in his previous country or the native who puts an unseemly past behind him and becomes an upstanding citizen, many believe in the American system and accept their place in it because they take it as an article of faith that anyone, no matter what they have done in the past, can start afresh and eventually succeed.

Reinvention is never quite as seamless as the myth implies, though, and Americans are rightly skeptical of those with tarnished pasts who claim to have turned over a new leaf. Apart from those segments of society with a long tradition of accepting people as born again after sufficient public shaming and atonement, a person's prospects for reinvention have always partially been a function of their ability to keep their past obscure enough to be plausibly benign. Data-mining makes recasting one's background impossible, however. You are no longer the resume you choose to hand to an interviewer or the persona you choose to present to a bank's loan officer. Your entire background and everything that can be inferred from it about your future prospects as an employee or debtor are available and beyond your control to frame in a positive light. When combined with the often unspoken skepticism that belies Americans' belief in individuals' capacity for change, this loss of control over the public perception of one's past may eliminate many wayward souls' chances for redemption.

Besides making it more difficult for individuals with self-imposed damaging pasts to move on, data-mining also harms those who must dig themselves out of backgrounds they did not choose to have. Generations of Americans have grown up believing that in our post-feudal society, debts reside and die with those who incurred them and are not passed on to their children. While technically still true, if it can be inferred that your parents' tendencies towards profligate spending was passed down to you, you perhaps may as well have inherited their debts if the result is higher fees and interest rates thanks to an omniscient credit rating agency. And ironically for a country in which there is so much bloviation about ``family values,'' those who will be most negatively affected are those who are deemed most likely to be loyal to and supportive of their kin because it will be inferred that they will come to their relatives' aid even at the risk of their own finances.

In the same way, data-mining may serve to reinforce the hardships endured by those who try to extricate themselves from backgrounds determined by race and class. Up until recently, a person ``of the other color[s]'' born in a neighborhood that had been red-lined could still, if sufficiently committed to changing their identity, have had a decent chance at home-ownership. By moving to a new, hopefully more tolerant area and ensuring that they dressed, spoke and acted the ``right'' way, they could convince a loan officer that they were not like ``the others.'' The same was largely true for attempts to enter other parts of the upper echelons of American society. There is no escaping the ghetto in the era of data-mining, however. If someone with power chooses to look hard enough, either blatant bigotry or well-meaning but invidious calculations may eliminate what was an, albeit demeaning for some, effective method of advancing in American society.

But since data-mining doesn't seem to be going away anytime soon, it may be wise to consider whether its killing of the American myth of the possibility of reinvention, and the related myth of the self-made man, has a silver lining. While Gladwell (and his editor) must have been overjoyed to learn that so many Americans were surprised to read that one's outcome in life is largely a product of one's upbringing and surroundings, the fact that it was such a revelation shows that the American myth of the self-made man--one who can succeed regardless of growing up under less than promising circumstances--may have been too powerful. While today Americans of all classes vigorously reject the notion that we are restrained by our backgrounds, if enough inference-making by the purchasers of our data makes this inevitably true, perhaps we will all have to rexamine our myths and think much more broadly about what must be done to make America a true meritocracy.


I think you touch on two of the significant tangible risks of data mining (discrimination and limitations on social mobility). I suppose abuse of data and errors in the data and process would be other risks. What I find more difficult to articulate are the abstract risks of data mining - this instinct that it's just inherently wrong to assemble and aggregate data about individuals. Although if that instinct has something to do with free will, that ties back into the point of your paper nicely.

-- ElizabethDoisy - 27 May 2009

Hey Ward, interesting paper - I think liz touched on two points that you may want to address more directly (if you think they make sense): the first is the free will point, that I think even goes beyond what you mention in your paper. A world where the powers that be know your every action and response would be more efficient, but at some point becomes indistinguishable from a finely oiled and well crafted machine. While economics and efficiency can aid us in attaining our goals, they are a means, not an end. Even if knowledge could get us to a capitalist utopia (which I doubt, as self interest seems often to work at odds with "efficient markets"), a world of purely efficient production and consumption makes no intuitive sense as necessary or even desired end for human civilization. I like the analysis in your paper, but I think the 1L was fundamentally operating from a flawed assumption: that more efficient markets are some kind of panacea for the human condition. Why would a rational person prefer his world over any other?

The second point is related, but is more philosophical: I think our humanity and empathy require ambiguity to function. Reducing people to a set of predicted behaviors robs them of their inherent value as human beings. Everyone you meet has as rich and worthy an inner life as your own - they are as human and as complex and interesting as you yourself. The more we reduce them to behaviors and models, the more we make it easy to forget or ignore their essential humanity. Mankind needs ambiguity - as soon as we subsume the spirit and endless complexity of our fellow man in a behavioral model, we lose the very thing that gives him worth, and destroy the reason we developed our fancy economic systems in the first place.

(also, "seemless" at the beginning of the second paragraph should be "seamless")

-- TheodoreSmith - 01 Jun 2009

Excellent point regarding the danger of elevating market efficiency from mere instrumentality for the pursuit of human happiness to an end in itself, Ted.

This is somewhat off-topic, but anyone who wants proof that we're heading towards a society that treats human beings as fully-modeled consumers on whom products can be 'pushed' with predictable results need only read this article in today's New York Times.

-- AndreiVoinigescu - 02 Jun 2009

  • I think this is an effective and valuable essay. Like your first, and in a very different way like the essays that Dana has been writing, like the responses to Matt Eckman, this is an attempt to broaden the cultural relevance of the issues for those who are not likely to find their way to them quite so easily as we. Naturally I believe that's very important intellectual activity with significant political and social consequences.

  • In reading what has been written this term, by you and your colleagues, I'm struck by the difficulty of asking people to imagine problems posed by long-term consequences when they are having a great deal of conceptual and logistical trouble just keeping up with what is already happening around them. Your recurrent task is to alert your reader to potential consequences that would ensue after many years of the further development of trends barely starting to be visible and which are not much reported or documented. Explaining concerns that depend on predictable but so far unexperienced developments is, as Kate says, the business of traditional Bradburyist science fiction. Maybe this is a genre worthy of our more serious attention.

  • The weak place in your present draft, it seems to me, is the conclusion. It is logically possible, but not either intellectually or emotionally satisfying, to conclude that the good thing about this technology's incompatibility with our founding social myth and operating commitment is that it may cause us not to put too much stress on them. Your own previously-expressed commitments seem to suggest the desirability of looking elsewhere for a conclusion, including the possibility of not so easily conceding defeat. It feels to me that we owe the symbolic structure of our free society more than just a quick but decent burial. That seems to be your argument. Though not quite to the end.

I agree that this is a very interesting paper. Although based on your title, I almost expected you to argue that Americans aren't frightened by data-mining driven determinations about their future because they come from a religious tradition of predetermination.

I think you need to directly address the issue of Free Will that Liz, Ted, and Andrei mention. When you're faced with the hypothetical you set out in the first paragraph, i.e. a world in which data mining results in hyper-accurate modeling of behavior, there are two possible outcomes. If there is free will, such modeling will never be accurate enough to justify the intrusions into our personal lives. If, on the other hand, such modeling is possible, it brings into serious doubt the very idea of free will, social mobility, and the self made man. That's the fulcrum of the issue right there. All of your retorts hinge on one assumption - that hyper-accurate modeling is not possible.

Otherwise you're left fighting for a principle that has no empirical backing. There's no point in arguing for exceptions to the rule when the rulemaker will be able to say with near perfect accuracy that no exceptional cases exist.

-- RazaPanjwani - 10 Aug 2009

 

Navigation

Webs Webs

r7 - 05 Jan 2010 - 22:33:58 - IanSullivan
This site is powered by the TWiki collaboration platform.
All material on this collaboration platform is the property of the contributing authors.
All material marked as authored by Eben Moglen is available under the license terms CC-BY-SA version 4.
Syndicate this site RSSATOM