A new strategic document calls on the US military to operate in a “wartime” format, adopt agent-based artificial intelligence at an unprecedented pace – and entrust autonomous systems with critical roles on the battlefield
I have read many strategic documents from the United States Army, but this is the first time I have been left shocked at the end of reading them. Last month's strategic document, launched directly by the Secretary of War, dictates a new direction for the entire army, and in the process makes it clear that we are entering a very different era.
Oh, and he also tells the generals that they need to recreate "Ender's Game" – the famous science fiction book – in real life.
But let's start at the beginning.
The United States Secretary of War sent an official letter in early January to all generals, commanders, and administrators in the army and the War Department. To summarize the letter briefly, it prepares the army for war. It calls on all commanders to abandon bureaucracy, procrastination, and long and tedious procedures. The War Department is entering a "wartime period," as he puts it.
“[A] wartime approach to impediments must be taken,” the document states. “We must eliminate impediments to information sharing, operational approvals, testing, evaluation… and other policy factors that hinder rapid testing and field launch. We must examine risk balances… as if we were at war.”
These are not words I would expect to see in an official document of the United States military. This is the most orderly, most ‘yuck’ and most pedantic military in the world. Everything is checked with tweezers. Every screw has its own name, with piles of approvals and standards behind it. And now, to win the artificial intelligence race, all this meticulous ethos is being swept aside. And why? For the third time – because we are at war.
So what do we do to win the war?
Artificial Intelligence as a Tool to Win Wars
Believe it or not, the answer is very simple: investing in artificial intelligence. Or more precisely, in the race to ensure that the United States military maintains dominance in the world of artificial intelligence. And as the Secretary of War writes –
“We must accept that military AI will be a race for the foreseeable future, and therefore speed wins. … We must accept that the risks of not moving fast enough outweigh the risks of an imperfect set [of AI values].”
And so, the entire document focuses on artificial intelligence, artificial intelligence, and some more artificial intelligence. The Secretary of War is demanding that the generals identify the most talented artificial intelligence talents, recruit them, and retain them within the military. He wants every new and sophisticated model that comes to market to be implemented in the military within just thirty days. Yes, you read that right: thirty days, just one month. In the United States military, until now, a month was the time it took to decide which new dish to add to the menu in the dining room. From now on, that will be the amount of time it takes for soldiers to develop sensitive and advanced artificial intelligence systems and implement them in weapons of all types.
As you can imagine, implementing such complex systems will not happen easily. There are plenty of bureaucratic, administrative, and logical barriers that will stand in the way. So what do we do? We will break through them, of course. The Minister of War orders the establishment of a "Committee for Removing Barriers," which will be able to wave aside any barrier and any resistance - just to allow things to happen quickly.
Because we are at war, remember?
Each unit, according to the guidelines, is supposed to have an "AI integration leader": a soldier who will be responsible for integrating artificial intelligence into the unit, and will push the issue with all his might. He should, because the Secretary of War oversees everything from above, and he wants to see results. And everything will be monitored, supervised, and controlled, with systems that "will collect feedback in days, instead of years."
Swarms, murderous agents and simulations
What will they do with all this artificial intelligence in the army? The document sets out a number of areas in which it will be integrated, and which it is supposed to catapult them forward. These include, for example, the "swarm reactor." This is a project that will research, discover and adopt innovative ways of fighting with and against artificial intelligence. Or in simpler words: learning to fight like the Ukrainians and against the Ukrainians (and the Russians) with swarms of robotic aircraft alongside the soldiers.
What else? Nothing major. Just developing AI agents that can fight alongside and in place of the generals. These agents will be able to take control of any task, "from campaign planning to executing a chain of executions."
And again, in simple terms: the Americans want to give artificial intelligence the reins on the battlefield. Yes, to the point of deciding who to kill and how.
A third interesting direction is explicitly and nominally based on the science fiction book "Ender's Game." The book, set in the future, describes a virtual reality simulation system that helps the hero of the story - Ender - learn about himself and understand himself better. The Minister of War gently asks his subordinates to set up similar simulation systems, which will help study the future battlefield before it even appears on the ground, and thus - "ensure that we stay ahead of artificial intelligence-based opponents."
Last but not least, there is no time or place for being nice in war. In today's America, this means that principles like "equality, diversity, and inclusion" are no longer relevant in the military. But the Secretary of War takes things a step further, stating that the artificial intelligences the military will use must also abandon these principles. The intelligences must provide "objectively correct" answers to the questions presented to them. But of course, what is this "objectivity"? Who determines what is "correct" and according to what criteria?
Now that you've read the demands of the Minister of War of the superpower with the most advanced military in the world, do you understand why I'm so anxious?
If not, then I will explain.
Reasons for concern
Let's start by saying that it is possible and desirable to integrate artificial intelligence into military decision-making processes. Yes, even in deciding who to shoot on the battlefield. Artificial intelligence that is properly calibrated and well-tested may also reach a level of morality that the ordinary soldier in the field is incapable of, when he is in the throes of battle and flooded with adrenaline. But to reach such a level of capabilities, artificial intelligence must be developed carefully and meticulously and tested in many different conditions in the field. Any mistake can cost human lives. Not only those of the United States' enemies, but also those of the soldiers themselves.
The Minister of War throws these concerns to the wind, and states facts on the ground: rushing forward at any cost. Or not at any cost, but at least at the cost of the lives of people who may fall victim to war machines that do not see well, hear well, or analyze the situation well.
In fact, even a correct analysis of the situation is not necessarily on the list of critical demands of the Secretary of War. The principles of “equality, diversity, and inclusion” were not established out of thin air. They came to provide a response to certain social needs, and to reflect high and worthy moral values that characterize liberal democracies. It is possible, of course, to take these values too far – and indeed many Trump supporters claim that this is the case in the United States right now – but it is not right to demand that artificial intelligences ignore them completely. And if that is what the Secretary of War demands, it is difficult for me to automatically assume that his war machines will be endowed with a developed conscience, or that they will obey the accepted laws of war.
If we were in the middle of World War II right now, with millions of soldiers and civilians dying on each side every month, I could understand the willingness to throw off any burden and demand rapid forward movement at any cost. But the United States is not waging a feverish war at this time, and it has no troops in Ukraine. So what is going on here? What war are they preparing for? Against whom? What does the honorable minister know that we don’t?
One could argue that the United States is preparing for war with Iran, with its tens of thousands of unmanned aerial vehicles. Or perhaps China, which has itself raced ahead with artificial intelligence, and recently released a video showing its impressive military robots. Or – who knows – maybe even Trump wants to go head-to-head with the Europeans over Greenland. The chances of that are slim, of course, but possible.
The only thing that can be said for sure is that we are in a time of great uncertainty. Europe, China, and the United States are preparing for war. Ukraine and Russia are already in the midst of it. And when so many guns are manufactured and laid out on the table in the first act, someone will use them by the end of the show.
And if we follow the teachings of the American Secretary of War, it is certainly possible that the ones who will use these guns will be the robots.
More of the topic in Hayadan:
