BY PETER HARRIS

Part 1 of 2 is available here

It is good to save the best till last, and in his analysis, Kissinger does just that. But if his analysis at this point is at its most original, it is also at its most terrifying as it concerns a novel threat that is existential in scope, but which can be understood analogously again through the First World War. There is, as the writer of Ecclesiastes concludes, nothing really new under the sun.

Kissinger raises the problem of artificial intelligence which has the potential to take out of the hands of human legislators the decision to go to war through the existence of autonomous weapons capable of defining threats and being able to start their own war. According to Kissinger, once such high-tech weaponry and its supporting systems become standard with computers assuming the role of principal determinants and executors of strategy, human input will be negligible.

War may therefore come whether humans want it or not.  

Kissinger does not object to such weapons per se. His justified concern is that there is no theory for understanding and controlling to human benefit a military world comprised of standardised AI weaponry and that such a theory may never happen because meaningful dialogue between nations and alliances cannot happen without the risk of disclosing new discoveries to potentially hostile powers. Yet, incorporating advanced weaponry into strategic thinking is as much a priority, Kissinger opines, as climate change.

As one at the end of his life, and like so many others in the same position, Kissinger was lucidly troubled by how the future might unfold. In his wise analysis, he identifies technical experts and historians as essential at shaping strategical theory for AI weapons. It is easy to see how technologists are indispensable to that process, but historians? The answer lies in the fact that it is historians who understand how the interface between technology and human bellicosity has shaped the causes and courses of wars over millennia. If the interface between AI and humanity in war-making is to be understood, historians are essential to it.

An example of this is from the war with which Kissinger began-the First World War. That War was the most technically advanced war fought to date, though its combatants had little idea of how much the technology they had would help turn the War into one of costly stalemate before recreating it as a war of movement of unprecedented speed in its final year. In terms of the catalysts for the War, there is something that is analogous to the present dangers of AI, though in a basic way. The industrialised nations of Europe in 1914 depended on trains to mobilise and deploy quickly large numbers of troops. This lesson had been learned by observers of the Franco-Prussian War when Helmuth von Moltke gained a strategical advantage by using the railway to concentrate 462,000 Prussian and allied German troops on the French border to face a French army of only 270,000. Therefore, once war was declared in 1914, troops were rushed to the front to gain the initiative, but that made war more certain as the railways had reduced the amount of time between the declaration and prosecution of war and thus minimised the time available for last-ditch peace initiatives. There was also something quite fatalistic about the situation as noted by the historian A J P Taylor who hypothesised that once the carriages and wagons began to roll, they did so “remorselessly and inevitably to their predestined goal.” In other words, each of the interlinking pieces of the railway plan smoothly and swiftly snapped into place once activated, therefore outpacing any attempt by anyone to interrupt the process. It was the inherent nature of rail technology’s complexity and speed that helped to catalyse the War and which shadows how the inherent autonomy and even swifter pace of AI weaponry not only will catalyse wars but constitute a casus belli.

Foreign policy experts, such as Jeremi Suri at the University of Texas, talk of Kissinger having created a structure for global stability. Suri characterises it as consisting of nuclear deterrence, the use of proxy war to extend American interests and détente with China through giving China access to the world’s markets. Kissinger’s support of Ukraine and of American help with materiel is consistent with his policy of proxy war; his support for a negotiated settlement is in line with his view that when proxy wars have served their purpose, they ought to be brought to a peaceful conclusion. Typical of how Kissinger managed his system was his view that Russia ought to be preserved as a great power because it has been and continues to be vital to global stability. What Kissinger could not have foreseen as he was building his system back in the late sixties and seventies was that the greatest threat to it was not humans but their invention of AI. If the nuclear deterrent keeps the peace between the great powers and proxy wars enable the US to pursue its interests without antagonising isolationist elements at home, there is no guarantee that autonomous AI weaponry without a strategic theory to nullify that autonomy will come to the conclusion that nuclear war is not worth it and proxy wars are preferable to direct clashes between the great powers.

It is to Kissinger’s credit that he saw this threat fully for what it is and began to address it. He not only regarded AI from the point of view of war, but co-authored a book about it with Google’s Erich Schmidt which concluded that AI was close to supplanting human control of the planet and was a challenge to human consciousness greater than anything since the Enlightenment. Kissinger’s legacy to the world may well be the international architecture he established and worked hard to sustain, but it is more importantly his warnings about AI whose revolutionary capacities he foresaw as capable of exploding that architecture but will now never experience himself.    

Peter Harris is the author of two books, The Rage Against the Light: Why Christopher Hitchens Was Wrong (2019) and Do You Believe It? A Guide to a Reasonable Christian Faith (2020).