Automatic for the people

By Tiago Mata

Three days after the inauguration of Donald Trump, the New Yorker published a profile of the Social and Behavioral Sciences Team, a small office created by Barack Obama in September 2015. In “Good Behavior” the reader accompanies Maya Shankar, the leader of the team, journeying to Flint, Michigan to aid a community injured by neglect and collapsing infrastructure. It was the New Yorker‘s farewell tribute to the Obama Presidency. Predictably, the team of behavioral scientists was an early candidate for Trump’s “you’re fired,” and their website now bears the disclaimer that its contents are “historical material ‘frozen in time’ on January 20, 2017.” In the profile, Shankar stands in for Obama. She is young, intelligent, accomplished, relaxed, devoted, at one point pledging to go without sleep to serve the people of Flint as the clock ticks towards her inevitable eviction from the West Wing.

Making Shankar and Conway into emblems of Presidency lost and gained binds us to misleading binaries: fact vs. alt-fact, expert vs. demagogue, reason vs. affect. We miss the continuities

The departure of Shankar appeared in print as a new White House occupant, Kellyanne Conway, asserted “alternative facts” about the Trump inauguration. In hindsight it is remarkable that we all cared so deeply about counting the heads under the drizzle of the National Mall. At the time, it seemed a fundamental insight on the character of the Trump White House, and it was a civic duty to explain that profaning matters of fact is the hallmark of totalitarianism. George Orwell would have been only 33 years off target when he foresaw that “the past was erased, the erasure was forgotten, the lie became the truth.” The sinister scenario of the coming of fascism seems less plausible today, after three months living in the disruptive and baffling age of Trump.

However serendipitous, making Shankar and Conway into emblems of Presidency lost and gained binds us to misleading binaries: fact vs. alt-fact, expert vs. demagogue, reason vs. affect. We miss the continuities. The 2017 winner of the prize for “Big Data” from The Advertising Research Foundation was Cambridge Analytica. Basking in the glow of Trump’s election, the firm bragged of amassing profiles of 200 million voters and using behavioral science models to nudge votes for Trump. Steve Bannon sits on its board. Even if the firm’s claims are exaggerated, we are discovering that the Democratic Party does not hold the patent for “nudging.”

Within the Obama administration, Cass Sunstein was the self-appointed champion for blending big data and behavioral science and he aided Shankar in setting up the behavioral sciences team. In self-defense but also alarm, Sunstein has written on the ethics of influence. His rules for socially responsible nudging prescribe that the ends of nudges must be legitimate and that nudges must be transparent and subject to public scrutiny. This light ethical burden is unlikely to counter the popular suspicion that nudging is a form of thought control, much like “brainwashing.”

I shall not dwell on the boundary work of good meets bad nudging and lose myself in another partisan binary. I rather note on what Obama’s policy makers and the bipartisan electoral marketers share, a critique of evidence and of expertise taken from behavioral economics. In a recent biography of Amos Tversky and Daniel Kahneman, Michael Lewis records how their quirky experiments targeted the expert claims of Israeli generals, tenured professors and clinicians. To behavioral economics, all experts are equal before the law of human cognition. They are biased in cognition and in decision-making.

One salient response to the undoing of epistemic authority has been a fascination for a new kind of public numbers. The old variety were credentialed statistics, such as headcounts on the National Mall, unemployment rates, global temperature estimates. These numbers travel tied to those that issue them, and we are routinely instructed to notice the biases of the interests and the biases of the brains of these experts and institutions. Similarly doubted is the interpretative evidence taken from focus groups, polls and surveys where participants are said to lie and “herd,” a prospect that is set to bring about a crisis of empirical sociology. The new generation of public numbers are the numerical waste of digital living. The quantified self of centuries past came into being through ideals of moral accounting and personal improvement which motivated individuals to make markings on diaries and balance sheets. The 21st century quantified self is recorded by algorithm, silently, unwittingly, automatically. Facebook, Amazon, Google, transport authorities, and the U.S. Government are doing it for us. Big data and behavioral science travel hand in hand and in your phone. If the New Yorker had accompanied Maya Shankar on inauguration day it would have followed her to the campus of Google, where she is the new Head of Behavioral Science.

The behavioral critique of expertise is congenial to populists, to digital entrepreneurs, and to the pro-business lobby seeking to overhaul regulatory agencies under Trump. The enduring attraction of nudging lies in its erasing of the problem of adjudicating expertise. In place of the thick analysis of institutional cultures that science studies provides, we have simple-minded claims about human cognition, harnessed from experiment or machine learning. In place of enriching venues of deliberation, we have promises of choice architectures that automate welfare, leading to choices of healthier diets, affordable insurance, and commitment to a pension plan. Faced with the spectacle of Trump’s oversized pride in his own judgment, we may soon all be yearning for automatic government.

Tiago Mata is a Lecturer in the Department of Science and Technology Studies at University College London.