Market Mad House

In individuals, insanity is rare; but in groups, parties, nations and epochs, it is the rule. Friedrich Nietzsche

My Thoughts

Will Artificial Intelligence Make the Thucydides Trap Worse?

Artificial intelligence (AI) could turn the Thucydides Trap into a state of permanent warfare between the great powers.

To clarify, the Thucydides Trap is foreign policy expert Graham Allison’s description for a conflict between a rising power and an established great power or superpower. In particular, Allison believes the United States and the People’s Republic of China could be caught in the Thucydides Trap.

To elaborate the term Thucydides Trap comes from the ancient Greek historian Thucydides who wrote a history of the Peloponnesian War. In that conflict the two major Greek states; Athens and Sparta, destroyed each other in a war that that ended Greece’s Golden Age.

Essentially, Sparta was the established power which dominated the Greek world. Meanwhile, Athens was the arrogant new kid on the block out to overturn the established order. Instead, of settling the dispute, the war weakened Athens and Sparta; making Greece easy pickings for the father and son conquerors Philip of Macedon and Alexander the Great.

Did the Thucydides Trap make World War I Inevitable?

“It was the rise of Athens; and the fear that this inspired in Sparta, that made war inevitable,” Thucydides wrote of the Peloponnesian War.

Allison studied 16 historical conflicts he labels Thucydides Traps and found the great powers avoided war in just four of them. In particular, Allison thinks a Thucydides Trap led to World War I.

To explain, fear of a rising power; Germany, drove three established powers; the British and Russian empires and France, into an alliance. In response, Germany established its own alliance with three weaker powers; the Ottoman Empire (Turkey), the Austro-Hungarian Empire, and Italy for protection.

War erupted when Austria-Hungarian attacked Russia’s client state Serbia in retaliation for the assassination of the Archduke Franz Ferdinand’ the heir to the Imperial throne. In response, Germany, France, and Russia mobilized to support their allies. Within three years all the world’s major powers; including the United States, China, and the Japanese Empire, entered the conflict.

New technologies; including heavy artillery, machine guns, stable explosives, poison gas, airplanes, tanks, and submarines made World War I into one of the worst bloodbaths in human history. One reason why World War I was so destructive was that nobody understood the new military technology or its implications.

For instance, no military authority was expecting trench warfare when the troops marched off in 1914. In addition, experts dismissed air, tank, and submarine warfare as science fiction in 1914, just as observers dismiss AI warfare as science fiction today.

The 21st Century Thucydides Trap and Artificial Intelligence

Today’s world has some disturbing similarities to the world of 1914. In particular, the world’s two greatest powers, the United States and the People’s Republic of China, could be in a Thucydides Trap.

In addition, there is a weak declining power, the Russian Federation, trying to preserve its great power status by manipulating America and China. Like Austria-Hungary’s leaders in 1914, Vladimir Putin could believe his Chinese alliance will protect him from America. World War I, however, destroyed the Austro-Hungarian Empire.

Moreover, today’s great powers are deploying dangerous and poorly understood new military technologies, as in 1914. In today’s world; though, the most destructive of those technologies could be artificial intelligence.

The AI Arms Race has begun

Specifically, defense expert Dimitri Simes Jr. thinks the world’s major powers are in the middle of an “AI Arms Race.”

Russia; for instance, tested the Uran-9 robot tank, in combat in Syria, Task & Purpose reveals. The Uran-9 is a real life terminator equipped with missiles, machine guns, and automatic canons.

Furthermore, China’s People’s Liberation Army (PLA) could deploy AI-operated submarines in a few years, The South China Morning Post claims.  If they work as advertised, the AI subs could sink ships, perform reconnaissance, gather intelligence, and lay mines without putting human sailors at risk.

Interestingly, the submarines are one part of an ambitious Chinese plan to use AI to enhance the PLA’s naval power. Besides AI subs, the PLA is planning AI systems to assist human sub commanders and AI-operated surface warships.

Swarm Warfare is here

The British military is experimenting with a potentially more disruptive weapon; swarms of small drones that could overwhelm planes, tanks, infantry, and even ships, the BBC reports. Swarming drones with built-in bombs could attach themselves to a tank or warship and explode, for instance.

AI-piloted drone swarms could be a cheap and easy to deploy, weapons system weaker countries; such as the United Kingdom, could use to neutralize larger powers’ forces. In the next war, swarms of flying bombs could attack American infantry and military convoys.

Robot swarms already exist. Notably, the British tech company Ocado Group PLC (LON: OCDO) operates warehouses where swarms of robots pull and pack grocery orders. Not coincidentally, Her Majesty’s military is pioneering the militarization of robot swarms.

How AI could make the Thucydides Trap worse

I can imagine several scenarios where AI could make the Thucydides Trap worse.

First, what happens if a Chinese AI sub and an American AI sub encounter each other in the open sea? If something goes wrong, an American or Chinese AI could start World War III with no human input. Instead, the presidents of both nations could learn war has begun with a phone call from a confused admiral.

Second, the same scenario could play out on the ground with AI-driven tanks, or in the air with AI-piloted drones. For example, both Russia and the United States could deploy AI tanks to a Middle Eastern battlefield to help their allies on the ground. The tanks could trigger the Trap by shooting at each other.

Third, AI-operated weapons could make it easier for leaders to engage in a reckless intervention that could lead to war. To clarify, intervention will risk no soldiers’ lives. Instead, each side could deploy more and more AI weapons to counter the enemies.

Will Artificial Intelligence Lead to Endless War?

Fourth, artificial intelligence could lead to endless war by enabling leaders to fight enemies while taking no human casualties on their side.  

American critics like the Quincy Institute are already complaining about a “forever war” in which conflicts never end, because costs and risks are low. United States forces have been fighting in Afghanistan for nearly 18 years, for example.

The American War in Afghanistan is dragging on because of the low number of casualties, which creates no problems for politicians back home. AI could lengthen the forever war by allowing leaders to wage war with no human soldiers.

Fifth, AI-operated weapons could make it easy for developed nations, organized crime, terrorists, or corporations, to wage war on people in undeveloped countries. An oil company, or drug cartel, could deploy robot tanks; or drone swarms, against tribal people who object to its operations, for example. In such a conflict, peasants armed with shotguns; or old AK-47s, could have no defense against heavily armored AI-operated terminators.

Could AI Warfare Lead to Mass Casualties?

In addition, AI tanks, robots, or drones could slaughter hundreds or thousands of people with no political repercussions. Disturbingly, the many deaths caused by US drone attacks have not affected the American political discourse.

Frighteningly, Airwars claims  in Middle East air attacks killed 51,288 people between 2014 and 2019. If Airwars’ claim is true, the number of people killed in Middle East air attacks in the past five years, exceeds the number of deaths caused by the Nazi Blitz of England in 1940 and 1941. The Encyclopedia Britanica estimates the number of deaths in the Blitz at 43,000.

Significantly, there has no public outcry against the Middle Eastern air attacks in any of the countries staging them. The United States, Russia, France, the United Kingdom, Israel, and Saudi Arabia are all staging regular air attacks that kill civilians and nobody notices.

Under these circumstances, a future war in which AI-operated tanks, artillery, drones, and walking robots slaughter thousands while laying waste to a country is possible. A true nightmare will be thousands of innocent civilians caught in the crossfire of an AI battlefield.

Will AI Make Terrorism Worse?

Other nightmares include terrorists turning drone swarms, or AI-piloted submarines, against civilian targets. A frightening scenario is hundreds of drones packed with explosives flying into a city like London or Manhattan.

Remember, Al Qaeda brought America to its knees with just four suicide planes on September 11, 2001. A future Bin Laden could send swarms of hundreds of drones against his enemies.

Another scary event is an AI-submarine torpedoing, or shelling, giant cruise ships at sea. For instance, the world’s largest cruise ship; The Symphony of the Seas, can carry up to 6,680 passengers, Business Insider reports. Such floating pleasure palaces are logical targets for the high-tech terrorist of the future’s AI-operated weapons.  

An even more inviting target for drone swarms will be aircraft carriers such as Britain’s $5.02 billion Queen Elizabeth. Interestingly, a drone reportedly penetrated the Queen Elizabeth’s security and made an undetected landing on the carrier’s deck in 2017. The Queen Elizabeth is a perfect target for drones because she carries 40 warplanes and 1,600 sailors.

Therefore, another nightmare is a future war in which terrorists use AI to attack targets in developed countries and kill thousands. Consequently, major powers retaliate with mass attacks by AI-powered weapons that kill hundreds of thousands.

World War I proves terrorism can trigger all-out war. Remember, it was a Serbian terrorist attack on Austro-Hungary that triggered World War I. Just imagine the hell modern Britain could unleash if terrorists killed Prince William.

Is it Time to Ban AI Weapons?

21st Century Technology could make an ancient nightmare the Thucydides Trap into a modern horror. Perhaps it is time for the world’s leaders to try to prevent that horror by banning AI weapons.

The Chemical Weapons Convention; which bans chemical weapons and mandates their destruction, could be a model for an effort to ban AI weapons. Moreover we could expand the Organization for the Prohibition of Chemical Weapons (OPCW) to suppress AI weapons research.

We need to consider an AI weapons ban because artificial intelligence is already beating human champions at a wide variety of games including; StarCraft II, Go, Poker, and Chess. Therefore, human soldiers could stand little chance in battle against AI-directed weapons. The AI could anticipate all the human combatants’ moves and counter them, leaving people defenseless.

Frighteningly, governments are using open-sourced artificial intelligence tools for roles like surveillance. For instance, The Financial Times reports Britain’s surveillance agency; the Government Communications Headquarters  (GCHQ), used Alphabet’s (NASDAQ: GOOG) DeepMind AI as the basis for a smart surveillance system.

DeepMind can beat humans at strategy games like Go and StarCraft so it could theoretically outwit humans on the battlefield. Additionally, Alphabet’s (NASDAQ: GOOGL) Waymo subsidiary is using DeepMind to make self-driving vehicles more accurate, Electronic Design reports. Therefore, DeepMind could make weapons more accurate and deadlier.

My suspicion is that weapons designers are already using DeepMind to make to make armaments deadlier. Notably, the artificial intelligence research outfit OpenAI is keeping some of its algorithms secret to prevent the tech from falling into the wrong hands, The Financial Times claims.

We need to think about the dangers of AI weapons now before the Thucydides Trap makes an Artificial Intelligence War a frightening reality.

Further Reading:

AI Superpowers: China, Silicon Valley, and the New World Order

Kai-Fu Lee

Destined for War: Can America and China Escape Thucydides’s Trap?

Graham Allison

History of the Peloponnesian War