No-one pretends that democracy is perfect or all-wise," Winston Churchill observed in 1947. "Indeed, it has been said that democracy is the worst form of Government except all those other forms that have been tried from time to time." And most people tend to agree with Churchill's sentiment that nothing beats the wisdom of the crowd. In 2007, EU polls found that around the world, regardless of country, continent, age, gender or religion, about 80 percent of respondents believed democracy was the best way to run a society.
And yet, very few people felt this way until very recently. Throughout most of recorded history, democracy has consistently equated to mob rule. Even in November 1787, a mere two months after playing a leading role in drafting the U.S. Constitution, James Madison took it for granted that "democracies have ever been spectacles of turbulence and contention."
To me, this raises one of the biggest but least asked questions in global politics: Should we assume that we are cleverer than our predecessors and that we have finally figured out the best way of organizing communities, regardless of their circumstances? Or should we assume that because democracy has a history, it — like everything else in history — will someday pass away?
The Technological Challenge to Democracy
I have been thinking about this question a lot since this spring, when I attended a dinner hosted by Stanford University's cybersecurity committee on how cybersocial networks might affect democracy. (Serving on committees is generally the worst part of an academic's life, but some can be rather rewarding by exposing their members to a range of new ideas.) My fellow diners were mostly members of Stanford's political and computer science departments, and much of the conversation centered on the details of designing better voting machines or networking citizens for online town hall meetings. But some of the talk rose above the minutiae and into grander, more abstract speculation on what supercomputers and our growing ability to crunch large data sets might mean for the voice of the people.
The past decade has seen huge advances in algorithms for aggregating and identifying preferences. We have all become familiar with one of the more obvious consequences of this trend: Personalized ads, tailored to our individual browsing histories, pop up unbidden on our computer screens on a regular basis. But research has moved beyond studying its effects on decisions about consumption and on to decisions about justice and politics. When tested against the 68,000 individual votes that have been cast in the U.S. Supreme Court since 1953, a computer model known as {Marshall}+ correctly predicted the outcomes of 71 percent of cases. Meanwhile, an algorithm designed at University of Warwick predicted the outcome of the 2015 British election with 91 percent accuracy by analyzing political tweets.
Admittedly, there is still a long way to go. {Marshall}+ is good, but it cannot compete with Jacob Berlove, a 30-year-old resident of Queens, N.Y., who has no formal legal or statistical training but is able to predict the Supreme Court's votes with 80 percent accuracy. And while the Twitter-analyzing algorithm got many things right about the British vote, it was utterly wrong on the one thing that really mattered: the relative tallies of the Labor and Conservative parties.
Explanation:
