Why Democracy vs. Autocracy Misses the Point

May 3, 2022

Don't miss Professor Guéhenno's Artificial Intelligence & Equality podcast on this topic with Senior Fellow Anja Kaspersen.

I have always been a contrarian. I was a contrarian in 1989 when I wrote my first book, criticizing the idea—then widely held—that democracy had triumphed once and for all. And today I find that I’m a contrarian again with my new book, because everybody is talking about the confrontation between democracies and autocracies and I think that’s missing the point.

Something much more important is happening: the revolution of data, the Internet, and artificial intelligence. I believe we are on the cusp of an earthquake in the history of humanity of a kind that happens only once in hundreds of years. The most recent comparison is the Renaissance, and the pace of change today is much quicker than back then.

The institutions we built in the pre-data age are soon going to be completely overwhelmed, and thinking in terms of the old categories of democracies versus autocracies misses all the new challenges that they will have to face. This is a time of great peril as well as great promise, as was the Renaissance—not only the era of Leonard da Vinci, but also a century of religious wars.

The current revolution of data and algorithms is redistributing power in a way that cannot be compared to any historical shift. Traditionally we think of power concentrating in the hands of the leaders of states or big industrial companies. But power, increasingly, is in the hands of algorithms that are tasked (initially by humans) with learning and changing themselves, and evolve in ways we do not predict.

That means the owners of Google or Facebook or Amazon are not the masters of our destiny in the same sense as previous corporate titans. Similarly, while it is true to some extent that data will give dictators unprecedented power to manipulate society, they may also come to be dominated by the evolution of the algorithms on which they depend.

We see already how algorithms are reshaping politics. Social media has created self-contained tribes which do not speak to each other. The most important thing in democracy is not the vote itself, but the process of deliberation before the vote, and social media is quickly fragmenting the common ground on which such deliberations have been built.

How can societies exert control over how algorithms manage data, and whether they foster hatred or harmony? Institutions that are able to control this new power are not yet really in place. What they should look like will be one of the great debates of the future.

I don’t have the answers: I believe no human mind can anticipate the extent of the transformations that are going to happen. Indeed, I think the very notion that you can know today what will be the right institutions for the future is hubristic. The best institutions (and people) will be those that are most adaptable.

However, I believe that one promising approach is to think in terms of the relationship between the logic of knowledge and the logic of democracy. Take central banks as an example. The average citizen does not have a clue about how monetary policy works. Instead we rely on politicians to task the experts at central banks to try achieve a certain goal—it could be full employment, or a stable currency.

The pandemic provides another example. Politicians often said they were following expert advice from panels of scientists. But the role of the scientists should be limited. They are experts in the risks and harms of the virus—but policy choices to respond to the virus all came with other kinds of risks and harms, and weighing them up is essentially a political decision, not a scientific one.

These examples illuminate how we should think about controlling the power of data. The average person has no clue how algorithms work, so we need the expertise of specialists. But how to balance between competing values that define a society is a political decision. We need politicians who do not claim to know better than scientists, and scientists who do not try to dictate policy.

We must also accept that different societies will have a different balance between competing values, such as collectivism and individualism. We look today at Russia pushing its ethno-nationalist agenda, and congratulate ourselves that we are the good guys. But can we define what, fundamentally, holds our own societies together? "Not ethno-nationalism" is a good start, but only the first step.

There will not be a single global answer, and we should not aspire to find one. Instead we need a plurality of responses, to give us a better chance of seeing what works and what doesn’t. Autocracy versus democracy may seem like the question of the day, but the coming revolution will rapidly force us to think in more novel and nuanced ways.

Jean-Marie Guéhenno is Arnold A. Saltzman Professor of Political Practice at Columbia University, an advisory board member of Carnegie Council’s Artificial Intelligence & Equality Initiative (AIEI), and former United Nations under-secretary-general for peacekeeping operations.

You may also like

NOV 10, 2021 Article

Why Are We Failing at the Ethics of AI?

As you read this, AI systems and algorithmic technologies are being embedded and scaled far more quickly than existing governance frameworks (i.e., the rules ...

JUN 15, 2022 Article

Deus ex machina or Diabolus ex machina?

"Whenever we see or hear something that resembles "deus ex machina," or an inexplicable use of technology that resolves itself too simply or easily, we ...

MAY 10, 2022 Podcast

Making Decisions When Values Conflict or Are Prioritized Differently, with Paul Root Wolpe

In this Artificial Intelligence & Equality podcast, Carnegie-Uehiro Fellow Wendell Wallach sits down with Emory University's Professor Paul Root Wolpe for a thought-provoking conversation about the ...