Big Data Democracy

Among recaps of President Obama’s 2012 re-election, this piece, on the analytics and modeling done by the Obama campaign in rooms known as the Cave and the Alley, caught my eye.  And creeped me out.  Here are some very speculative and possibly provocative thoughts on “Big Data Democracy,” in no particular order:

There was a time when public policy was meant to be determined by reasoned debate about the welfare of the people.  That’s the high school civics model of democracy, at any rate.  If it is was ever accurate, it isn’t any longer, and not just because of the virulent, personal character of modern political advocacy.  The background assumption that we are reasoning, information-processing decisionmakers is increasingly being replaced by the sum of several dozen behavioral and demographic characteristics.

Public life, and distinctions between public life and private life, depend mostly on some long-standing and well-ordered philosophical assumptions about relationships among “being,” “choosing,” “thinking,” and “believing.” Those relationships are ever more unsettled as a psychological and technological matter. The unsettling is due in equal parts to the success of predictive models of the sort used by the Obama campaign, to Big Data’s computational power, and to access to the individualized and/or private information that feeds the Big Data machine. Neuroscience research and the insights of psychologists of various stripes have a lot to do with it as well, but it’s taken the quants and the geeks to get those insights to scale.

The secrecy of the whole venture, while completely understandable from the campaign’s perspective, reminds me of corporate efforts to maintain the confidentiality and security of customer lists, marketing strategies, and chemical formulas. Our democracy now appears to be the product, in part, of non-commercial trade secrets. I can only imagine the kind of dirty-tricks-as-corporate-espionage that may ensue.

Some of all of this is not truly new. I was a campaign organizer in the 1983-1984 Democratic Party presidential caucus in Iowa, and each of the campaigns back then had access to a Party database of caucus attendance from the 1980 and 1976 cycles. We marked Kennedy supporters and Carter supporters as likely supporters of different candidates in ’84. (The Democratic Party caucus in Iowa is not a public election, and expressions of preference by caucus attendees are not private.  You have to stand in a corner, literally or metaphorically, in view of the rest of the caucus, to stand for your candidates. The Party recorded – and may still record – who stood where.)  When I have described that experience to friends over the years, most have been pretty surprised by the existence and use of what they assume is private information. But its granularity and scale is dwarfed by what is possible now. So new it’s not, but its power is different. I think, along with lots of other observers of that last week, that it’s a whole new ballgame.

What’s the takeaway?

I wonder, when all is said and done, whether we should continue to think about the sources and uses of power and authority, about democratic legitimacy, and about the rule of law as we have done for the last few hundred years. Do we need a Federalist Papers for the 21st century? Literally and/or metaphorically? One that takes account of modern conceptions of information privacy, of the role of opportunities and challenges associated with secrecy in public life, of the power and peril of modern IT, and that builds upward from a model of human behavior that doesn’t depend, in the first place, on an “information processing” model of human behavior that assumes that we decide after all of the facts and opinions are known?