Opinion: Computational politics

Unlike all-powerful individuals, computers can be made unfeeling to the allure of tyranny.  (Thomas Galvez/Flickr, Creative Commons)

Unlike all-powerful individuals, computers can be made unfeeling to the allure of tyranny. (Thomas Galvez/Flickr, Creative Commons)

Since Plato established the notion of a philosopher-king, a trained individual who can weigh the consequences of his actions with far greater ability than military men and bureaucrats, the will of the masses and the all-powerful untrained individual as reasonable compass for policy have been put in doubt. Time and time again, tyrants seize the helm of governance, serving no other but themselves, and the passions of the public run amuck, allowing these tyrants to rule, sometimes by mob. The Framers of the Constitution, in bringing about a republic, held at bay direct rule by the people, and in the separation of powers and federalism, sought to hold back surges of majoritarian sentiment.

Even in periods of relative calm, Christopher Achen and Larry Bartel, renowned political scientists and authors of Democracy for Realists, present substantial evidence collected over the past century that electorates in democracies around the world vote predictably based on group identity and economic conditions right around elections. This is not a system conducive to the improvement of political character, contrary to the foundational assumptions of democratic rule.

Not only do citizens not have the time or motivation to research the candidates, but these considerations are colored by group mentality and limited by the extent of personal knowledge. How can anyone besides a career lawmaker be expected to make a decision on economic policy, not to mention the multitude of other topics of which campaign platforms and the role of government consist?

Though well-crafted democracy is surely superior to power by singular rule, at least from the perspective of the citizenry, computers could hold the solution to incorporating the decisiveness inherent to authoritarianism into our democratic systems. Unlike all-powerful individuals, computers can be made unfeeling to the allure of tyranny. Unlike voting citizens, computers could be made to systematically analyze all available policy options. Here is the philosopher-king for which Plato yearned, guided by the general ideals and rights of citizens, but left to make the most rational possible decisions in all particular respects.

How could such a computer be implemented? One can imagine many “small” computers handling the more systematic duties of government, such as economic policy; there are enough data and formal constraints here to make plausible the efficacy of machine learning models. Chaotic systems such as foreign affairs could even be modeled given an effort to formalize them. But it will no doubt take a herculean effort to understand the many human factors at play, the results of political action, and how each action’s success may be weighed. Formal representation of dynamic systems is at the forefront of research in many scientific domains and will likely shed light on the complex systems of human interaction in the course of its progress.

A more measured approach may be to transfer the knowledge from existing experts to models ready to receive it, something like political apprenticeships. Even a balanced consideration of qualified perspectives will likely outperform any single individual (see ensemble learning), and will be able to more effectively represent collaboration than would a group of experts, who are often swayed by passions and the consensus of the social groups to which they subscribe rather than by uncut reasoning.

One could imagine the communication of these small computers increasing to the point of a single intelligence — an autocratic, yet deterministically fair and ultimately rational system of rule. If many nations adopt this governance, they will communicate through diplomatic channels more complete and capable of compromise than any human delegations, and may even converge to a global computational politics through humanity’s consent.

Of course, international aggression is open to increase as well, though this is a characteristic inherent to human nature, and a computer that would eliminate this sentiment must either rule over a pacifist people, or otherwise disobey its people and with it abandon its democratic tendencies. The risk of fighting meaningless wars, in any case, is likely to decrease, and the potential of nations will rest in the interests and character of its citizens.


Berk Alpay is a contributor for The Daily Campus and can be reached via email at berk.alpay@uconn.edu.