If artificial intelligence (AI) can predict future wars, can it also advise us on how such wars can be prevented, asks Professor Tshilidzi Marwala.
The Vice-Chancellor and Principal of the University of Johannesburg (UJ) as well as the author of the book Artificial Intelligence for Rational Decision Making, Prof Marwala recently penned an opinion piece, ‘Can artificial intelligence prevent future wars?‘, published by Sunday Independent, 06 May 2018.
Can artificial intelligence prevent future wars?
Will Lesotho ever go to war with Switzerland or South Africa ever go to war with Belarus or Lesotho fight South Africa? These are the questions that this article attempts to answer.
During the cold war, the Soviet Union and the (US) were involved in the ideological war in the Korean Peninsula. The result of this was that in 1948 South and North Korea became separate countries.
Subsequent to this, the two countries fought the Korean War between 1950 and 1953. The result of this war was the demilitarised zone that divided Korea since 1953 into the US-aligned South Korea, which was capitalist, and the Soviet-aligned North Korea, which was communist. Because of this division, South and North Korea are technically at war.
Last month, the South Korean leader, Moon Jae-in, and North Korean leader Kim Jong-un met to discuss the denuclearisation of the Korean peninsula and the de-escalation of this conflict. This was the first meeting of leaders of North and South Korea. Could we ever have predicted that in 2018 the leaders of North and South Korea would meet to end the technical war the two countries have engaged in for the past 65 years?
In 1989, when I was at Mbilwi Secondary School, I won the National Youth Science Olympiad and as a result, I attended the 1989 London International Youth Science Fortnight in England. At this forum, East and West Germany were two different countries and each had sent its own delegation. This was three weeks before the fall of the Berlin Wall.
Looking at the interactions of the delegations from these two German countries, I could not predict that just in three weeks the Berlin Wall would collapse and that the two nations would unite. Since I couldn’t predict the unification, could artificial intelligent (AI) machines have predicted this outcome?
Artificial intelligence is a technology that looks at how a human brain functions and designing a computer that is able to do what the human brain is able to do. As a result, we are able to create a machine that does well on tasks that humans are able to do, such as working in factories, listening, talking and thinking. These machines are intelligent, autonomous and are able to adapt and evolve beyond the specifications given by their human maker.
Bill Gates thinks AI is good for society while South Africa’s Elon Musk thinks AI is more dangerous than nuclear weapons. The late British physicist Steven Hawkins believed that AI would end mankind. A few weeks ago the President of the US, Donald Trump, used drones to bomb Syria. Drones are unmanned aircraft that use AI to navigate, identify a target and then bomb it.
They do these tasks without a human being inside them. Indeed in many ways, as observed by Musk and Hawkins, AI can be dangerous. But also in many ways, AI can be good for society. AI is able to diagnose diseases before they progress beyond control. It is able to predict the structural integrity of buildings and prevent their collapse and, thereby, save lives.
At the University of Johannesburg (UJ), Professor Thad Metz is working on AI and ethics. Mankind has successfully controlled destructive nuclear technology through the nuclear proliferation treaty and we should do the same for AI.
Going back to the question I posed at the beginning of this article, can AI tell us whether two countries will ever go to war and when? If it can predict future wars can it also advise us on how such wars can be prevented? Karl Marx believed that the driver of the history of society is conflict and that we should aim to change the world to eliminate it. If we use AI to eliminate future conflicts, what will then drive the wheel of history?
AI has been used to predict interstate conflicts. In our book, Militarized Modeling Using Computational Intelligence, we developed an AI machine that considered seven factors, also called variables that drive interstate conflicts. One of these factors is the distance between the two countries’ capitals. It turns out that the larger the distance between the capitals of two countries the less is the likelihood of a war between them.
The second variable that is important is whether two countries share a border or not. Countries that share a border are more likely to have conflict. The third factor is the level of trade between two countries.
The high the degree of trade between two countries the less likely is the probability of conflict between these two countries. The fourth factor is the relative difference in the degree of militarisation between two countries. If two countries are equally militarised the probability of war between them is reduced.
The higher the degree of militarisation of one country relative to another, the higher is the likelihood of war between these two countries. The fifth factor is the level of democracy in two countries. If both countries are democracies, then the likelihood of war is reduced. In fact, there is a theory in political science called the theory of democratic peace, which states that democracies never go to war.
The sixth factor is the number of agreements the two countries have signed. The higher the number of agreements between two countries the lower is the probability of conflict between them. The seventh criteria are whether at least one of the two countries is a superpower. Superpowers are able to fight distant wars. For example, the US was able to fight wars in Syria, Libya and Iraq while Russia was able to fight a war in Syria.
Now given all these factors, will Lesotho ever go to war with Switzerland? The distance between these two countries is 13 000km, they do not share a border and neither of them is superpowers. Using AI and all these seven variables, Lesotho and Switzerland will never go to war. The same AI process is used to analyse South Africa and Belarus, which are 14 000km apart, do not share a border and neither of them is global superpowers and again the outcome is that the two countries will never go to war.
As for Lesotho and South Africa, the distance between their capitals is 475km, Lesotho is completely enclosed, there are unequal levels of militarisation and South Africa is a regional superpower. It, therefore, was not odd that when there was a clash of interests in 1998, South Africa invaded Lesotho.
It is, therefore, imperative that as a society, we use AI technology to create more peace and prevent wars. As Albert Einstein put it: “I know not with what weapons World War III will be fought, but World War IV will be fought with sticks and stones” because any future world war will destroy civilisation and end history.
• The views expressed in this article are that of the author/s and do not necessarily reflect that of the University of Johannesburg.