Thoughts on Strategy, War and AI

Self-Reflection, Deception, and Degradation of "The People" (or political and strategic theory for a Sunday)

I suppose as someone who has taught “Strategy” to a number of graduate students over the years, I should have something to say about strategy and artificial intelligence (AI). And I do. But it might not be what you think. I want to take a longer view, one that current thinkers may be overlooking in their attempts to claim that AI is going to be “the” thing, capability, whatever, to winning wars, deterring wars or adversaries, or what-have-you. I certainly have many thoughts on the use of AI and autonomy in military operations. But let’s go up a level to strategy. Let’s start from an oft-cited premise: Clausewitz’s On War (1816-1830ish).

If war is “an act of force to compel our adversary to do our will” but on a “larger scale” than the classic duel, then war is nothing but the utilization of any means to manifest power over something or someone. This is what many refer to as the “nature” of war. This nature is unchanging, universal, and takes on the status of a Newtonian Law (1687).

While Clausewitz writes as if he wants to make his analysis of war on par with Newton, in many ways he is more following in the trends of the time, and in his Prussian backyard, and is more akin to proving some unified theory - a metaphysics, if you will - of war comparable to the works of Immanuel Kant. This is evident when he speaks of the “theory” of war versus the practical reality of its undertaking. Pure reason (1781) versus practical reason (1788). The Doctrine of Right (1797) versus a real world state with all of its legal, executive and legislative functions carried out in sometimes incoherent and inconsistent ways. It is also apparent when he makes claims about the necessary character traits for “military genius” - where discusses the need to follow “imperative maxim[s]” and such, as well as his discussions about the necessity of building a theory of war and its observations as merely “analytical” and “synthetical.” Such language is Kantian (1785). But I digress.

Because of the synthetical (or real world stuff), Clausewitz must make room for his “laws of probability.” Subjective analyses, estimations, beliefs, etc. all play their parts here, and it is why Clausewitz claims that “war is, therefore, not only a veritable chameleon, because in each concrete case it changes somewhat its character” but also because the universals cannot but manifest themselves in whatever is happening at any given point in time given the parties to war, their capabilities, their political goals, their values, their ideologies or even the winds and whims of Fortuna (to bring Machiavelli (1513/1532) in here).

So, where Clausewitz points to the “strange trinity” of Violence, Chance, & Policy (or political goals). They are represented then in his second trinity as the People, the Commander, and the Government. These points - say in three dimensional space/time - form a balance between them (maybe more Cartesian).

The importance of policy cannot be overstated for Clausewitz. War is a “continuation of politics by other means,” and thus it follows that “war belongs to policy” because “policy here is the representative of all interests of the whole community.” There can be faulty policies, to be sure, and this he implies is the root of the cause of conflict and war. But let us turn our attention to the famous Clausewitzian concept of the “center of gravity,” before we turn back to policy.

Centers of gravity are the “weights” of an enemy’s power. The are the grounding wires, so to speak. These centers of gravity may be manifold or few. They may be alliances, the troops, the economy, etc. But his advice is to attack these centers of gravity in “concentric lines” because “the concentric attack is, therefore, always that which may lead to greatest results.” Think of it like waves on a beach. Wave after wave.

One of these centers of gravity is also, of course, the people. If the people give the country/society the strength it requires, then it is obvious to break the ties that bind. It is interesting to note on a side-bar that when pushed, militaries always say that “their people” are their greatest strength. But those people come from somewhere…This is where I want to bring in the current discussions around AI, deception, manipulation, and the like.

I have written about AI deception before, where I argued that we need to understand the technology and its capabilities, as well as what is conceptually and practically required for “deceit.” I will not revisit that here. However, with renewed interest in the field of AI deception - where generative AI systems may be deployed to create “deepfakes” in video, audio, and any other sensory domain, there are renewed fears about how this deception may threaten “democracy” going forward. Most often, this is more about how AI is used in deceptive campaigns, psychological or information operations, etc. to influence elections. While these are all certainly very important issues, I want to take it one level up of abstraction.

What is important for democracies is not that they merely hold free and fair elections. Free and fair elections are the RESULT of having open, free, and democratic societies. Voting is really just casting one’s preference for a candidate. What is important is larger than that. Freedom of speech, freedom of the press, freedom of thought and action (within limits), self rule, separation of powers, universal suffrage (not for the act of preference but for the recognition of all citizens as equal). Citizens within the body politic are individual and equal parts of the whole. At least this is the ideal.

This view of the demos as a body politic is not new. Going back to the Greeks, we can look to Aristotle and his thoughts on what made up civic bonds. We needed to have something to “hold in common” (homonoia) with each other, to make up these ties. In democracies, we need to each freely assent to these ties and membership (if we follow Locke, Rousseau, to some extent Hobbes, and Kant). Social contract theory is about consent to being a member of the body politic, but also consent to be governed by laws and abide by those laws. But I digress.

The people, then, are the source of the strength in democracies. But it is not an individualistic notion were everyone is just summed up together (3 million vs. 5 million - thus 5 million is more and better). It is about the strength of the “we” in We the People (to be American about it).

So why is AI a threat to democracy? It is not because of election interference. That would be a threat to a particular election and its outcome. To be a threat to democracy, it needs to take aim at THE PEOPLE. So, enter Clausewitz.

If a belligerent wanted to break its adversary, it would go to its center of gravity, yes? If its center of gravity is the People, then it needs to find a way to break bonds of social trust, civic duty, civic engagement, and social ties with fellow citizens and its trust in government (and military) institutions. How would it do that? Military engagements actually tend to foster greater bonds of citizen engagement (if you are under attack and not doing the attacking, e.g. Rally Round’ the Flag). In the case of Western Liberal Democracies, they have very large and well equipped professional militaries, so fighting head on with them is risky and carries vast amounts of uncertainty. Plus, the military may not be the real center of gravity.

In all volunteer forces, it is often motives of civic duty that incentivize folks to enter into the armed forces. (There are also economic motives, and other motives too, but let’s not dwell here). So, if an adversary can break civic bonds, undermine trust in government institutions, create distrust amongst citizens or groups of citizens pitted against one another, then they have effectively made their adversary weak. There is little support for government, less support for the armed forces, foreign policy is affected, alliances are weakened or broken, and the individual member’s perception of the world is fundamentally changed.

The utilization of AI to manipulate beliefs, to change or warp perceptions, to “other” every thing and everyone who does not look, talk, believe, or act like you is the way to break the people. Fracturing a society through disinformation and manipulation is not new. Even during WWII all sides utilized deception and propaganda to achieve their aims. Wars since then have likewise utilized this same tactic. Indeed, during the Cold War, the former Soviet Union was a master at “Active Measures” ктивные мероприятия, or the ability to use disinformation to manipulate and deceive public opinion. The current Russian state is doing the exact same thing, using the exact same playbook. Putin is, after all, a Chekist, and he openly celebrates the longest serving KGB head and communist party leader (and Soviet leader) Yuri Andropov. There is even a holiday for Andropov.

China, well, same. But I’m not as steeped in this area of expertise, so I’ll leave that to the China experts.

The point is that to break the civic bonds, sow distrust, and manipulate beliefs has been the major focus against democracies for a really really long time. That part isn’t new. What is new is the EASE with which to do it. The SCALE that it can be done at, and the COST is minimal. Feedback loops within news cycles perpetuate it. Social network architectures and their optimization algorithms actively encourage it.

But it is also more than that. The ways in which we receive our information is continually modified and mediated through a screen. (Look, you are learning about some old dead white guys and some living white guys right now, reading it on a screen.) One has to work pretty hard these days (as opposed to yesteryear) to validate information coming your way. Breakdowns in civic bonds, again, is not new. Robert Putnam’s amazing work on civic and social bonds in Bowling Alone comes to mind. So you have a combinatorial effect.

What is the remedy, pray-tell? It is self-reflection, and what I would call “Re-Active Measures.” First, Western democracies, especially the US, needs to reflect on the breakages, cleavages, and social tensions that exist and have existed for *hundreds* of years. It needs to try to mend that social fabric. For if there is no “we” there is no “We the People,” but rather a haphazard collection of self-interested individuals going about their lives trying to maximize their own happiness, pleasure, wealth, position, reputation, or whatever, at the cost of everyone else around them. There is no bond - there is only transaction. That is a huge lift; I’m aware.

Second, there needs to be a greater awareness and education about how current AI systems and social media networking and platforms work. These are optimization models in many instances, so more eyeballs on, more attention, more data (for the companies). More time in front of that screen means less time out in the real world, living amongst and interacting with your fellow citizens. How do you do that? Well, “life hack” it. Find some form of behaviorist (i.e. BF Skinner) technique to condition yourself to be more neighborly. What is the operant that will get you there?

Third, read more. If you can’t travel due to time, commitments, or financial constraints, read more about other places and people. Read history, read autobiographies, check out a travel book (like Lonely Planet … pun intended) and read about a place you’ve wondered about. Learn another language. Learning another language means you can’t help but learn about another culture, their customs, values and beliefs too.

I’m a Kantian. Cards on the table. I wrote a book about his theories. He believed in a “Cosmopolitan” principle. That we should strive to engage with others because it 1) increased commerce but 2) lessened the likelihood for war. Kant had his flaws, sure. And no, he really didn’t think you couldn’t ever lie. He said there was no “Right” (Recht) to lie. But I digress, again.

The point is that if AI is going to be used as a strategic weapon - the means to manifest power over something or someone - then we should be prepared for its use against centers of gravity. In this case, the weapon doesn’t need to have a warhead. AI as a means to deliver the weapon of disinformation, misinformation, and active measures, will do more damage than any warhead. The Soviet Union didn’t crumble because of an all out WWIII. It went bankrupt. We are facing our own bankruptcy if we don’t reflect, educate, rebuild and take some re-active measures.