Computer Assisted Foreign Policy Decision Making

Foreign relations are complicated. The United States entered the war in Iraq with little knowledge of the many factions in that nation, and (in retrospect) obviously inadequate ability to predict how they would react to various actions of the occupying forces and of each other. Moreover, the Coalition forces should have understood the objectives and intentions of other nations as they affected the situation, and how they would evolve. It would also have been important to consider how the war and occupation would affect the internal affairs of many other nations, such as Lebanon, Saudi Arabia, the Emirates, Iran, and Turkey. The impact on the Israeli-Arab conflict should have been understood. Importantly, the different constituencies in the United States, the United Kingdom, and other allied and neutral countries should have been considered. How was the invasion and occupation going to affect U.S. diplomatic efforts in other spheres? How was it to affect the U.S. military, and the ability to exercise military power in other spheres were that to become necessary. What exactly were U.S. interests in the region?

Comparably, the decision to install anti-missile defenses in Central Europe similarly required complex considerations of the impact on the Russian Federation, which would involved considerations of the responses that would benefit the government in terms of its domestic constituencies, and its other foreign policy objectives. What would the EU countries think? How would the decision affect NATO? American foreign policy objectives in Iran, North Korea, and other regions of the world.

Ultimately, these decisions and many others are linked, and the United States requires a global foreign policy strategy, which depends on the global foreign policy strategies of other states, which in turn depend on the situations, goals, objectives, and resources of many peoples and factions worldwide.

Computers and Strategic Games


Chess is much simpler than foreign policy. There are sixteen pieces to a side, each with very limited numbers of moves at any time, on a 64 square board. After decades of effort, computers are now able to play chess at a world championship level. This was accomplished through the use of very powerful computers programmed very very complex programs developed through cutting edge efforts at artificial intelligence, and loaded with huge amounts of data on previous games and their outcomes.

Go is a more complicated game than chess, with 400 playing points on the board, each of which can be empty or occupied by either a black or a white stone. On the average there are more options for the player per move than in chess, and go games can have longer chains of moves than chess. Like chess, it is a game of strategy, depending on the control of spaces. Like chess, decades of effort have resulted in ever better computer systems for playing go. Go programs are not yet, I understand at the world champion level, but are now able to play at the level of very good players. (Like chess, go has a recognized system for evaluating the strength of tournament players.)

Can a Computer play foreign policy?

In the case of foreign policy, unfortunately, there is no recognized system for judging the quality of practitioners. Unfortunately, sometimes we put foreign policy in the hands of not very good policy makers.

Foreign policy is a very complex "game". There are 192 nation states, each with its domestic policy concerns that should be understood as well as its foreign policy. The variety of economic, security and other foreign policy concerns of each state should be understood, as well as those of non-state actors such as intergovernmental organizations, multinational corporations, and international civil society organizations. They are embedded not only in a very complex physical system undergoing environmental degradation, but in a complex institutional system in which international markets and other institutions are globalizing.

The question is not whether computers can master the "game" of foreign policy, but I suggest, how long it will be before there are computers that can play the game of foreign policy at a world class level. I think that will be a very long time! But the longer we wait to start developing such computer systems, the longer it will be until they exist.

In the case of chess playing computers, there was a war between human and machines. As machines got better, and able to beat many human players, the best human players found weaknesses in the computers that they could exploit to win. In turn, these informed the programmers, who developed still better computer programs while continuing to take advantages of advances in computer technology to build more powerful chess playing computers.

Today, the person learning chess would be well advised to play against machines, to develop skills and understanding of the game. I suspect that it would be very interesting to see computer assisted matches between the human chess grand masters. So too, long before we would turn over U.S. foreign policy to a computer, foreign policy computers could play a useful role in training diplomats, legislators, and elected officials. Indeed, I suspect our schools of foreign policy could learn about the field in the process of advancing the technology.

I did an internet search on the subject, and found relatively little research. Of course, that might be because the development of such systems is being done by intelligence agencies and ministries of foreign affairs behind a cloak of secrecy. Indeed, I hope that is the case!