When AI Decides Who to Strike: International Humanitarian Law Faces an Unprecedented Structural Shock

When Algorithms Decide 'Who to Strike'

Imagine this scenario: a drone circles above a battlefield. Its cameras capture ground targets. An AI system automatically classifies each as either an "enemy weapon" or a "civilian vehicle," then decides within milliseconds whether to open fire.

This is not science fiction. According to Opinio Juris, AI-enabled military target-selection systems are being rapidly deployed by armed forces around the world. The question is: when machines replace humans in deciding whether to fire, can the rules of war we have relied on for decades — the Geneva Conventions and their Additional Protocols — still hold?

Put simply, this is a head-on collision between technological logic and legal logic.

Two Core Safeguards Are Breaking Down

International Humanitarian Law (IHL) — the body of law governing conduct during armed conflict — rests on two key mechanisms designed to protect civilians:

• The Principle of Distinction: Parties to a conflict must always distinguish between combatants and civilians, and between military objectives and civilian objects. Indiscriminate attacks are prohibited.
• The Principle of Proportionality: Even when striking a legitimate military target, any incidental harm to civilians must not be "excessive in relation to the concrete and direct military advantage anticipated."

These principles sound reasonable. But the way AI systems operate is turning both safeguards into paper tigers.

The Distinction Dilemma: AI identifies targets through image recognition, signal signatures, and behavioral patterns. Yet the battlefield is an extraordinarily complex environment — a vehicle might be a troop transport or a family fleeing; a person might be carrying a hoe or a rocket launcher. AI's judgments depend on training data and probabilistic models. It lacks common sense, and it has no contextual understanding. When it misidentifies a target, the consequences are irreversible.

The Proportionality Dilemma: Proportionality requires weighing "military advantage" against "civilian harm." But how do you quantify military advantage? Who sets the "acceptable threshold" for civilian casualties? AI systems tend to speak in numbers, yet the cost of human life in war has never been a simple math problem.

Consider it this way: A human commander on the battlefield can hesitate, observe a little longer, or rely on intuition to call off a strike. An AI system, by contrast, prioritizes speed and efficiency — when "fast" becomes the top priority, "caution" becomes a system bug.

The Accountability Black Hole: Who Is Responsible When Things Go Wrong?

An even thornier issue is accountability.

In conventional warfare, when a strike goes wrong, the chain of responsibility is clear: the pilot, the commanding officer, and ultimately the state. But once an AI system enters the picture, that chain becomes blurred:

• Is it the fault of the algorithm engineer?
• The fault of the entity that provided the training data?
• The fault of the procurement decision-maker?
• Or is it simply "autonomous system decision" — and no one is responsible?

According to analysis by Opinio Juris, the existing international legal framework has no designated responsible party for "machine decision-making." This means that if an AI system causes mass civilian casualties, victims may face a situation where they do not even know whom to sue.

One way to read this: This is not just a legal issue — it is an ethical one. When life-and-death decisions are outsourced to algorithms, is humanity's moral responsibility being outsourced along with them?

Why Should Ordinary People Care?

You might think: "I'm not going to a battlefield — what does this have to do with me?"

It has everything to do with you.

First, the technology spillover effect. The AI used today for military target identification could be deployed tomorrow for urban surveillance, border control, or even social governance. If the logic of "algorithms deciding who should be targeted" becomes normalized, where do ordinary people's privacy and rights end?

Second, a global governance vacuum. International rules governing the military use of AI are virtually nonexistent. Nations are locked in a race to set the rules first, while civilian protection is sidelined. Without global consensus, an AI arms race could drag everyone into an even more dangerous situation.

Third, the escalation of information warfare. AI can do more than decide "who to strike" — it can generate false intelligence and fabricate battlefield situations. When information itself can be manipulated by algorithms, ordinary people's ability to access the truth is further eroded.

A note of caution: Some technology optimists argue that "AI is more precise than humans and can reduce collateral damage." But precision does not equal correctness — if the training data itself is biased, AI will simply make mistakes more efficiently.

What Can Ordinary People Do?

Faced with the wave of AI militarization, ordinary people are not powerless.

Stay informed and speak up: Support international organizations and initiatives pushing for transparency in the military use of AI. Public pressure is a powerful force for policy change.

Build digital literacy: Understand the basic logic of AI, and be wary of the myth of "algorithmic neutrality." Every algorithm reflects the values of its designers and the biases in its training data.

Advocate for responsible AI governance: Follow policy discussions on AI ethics and governance at both national and international levels. Use your vote and your consumer choices to express your position.

Broaden your perspective: Historically, every major technological shift — nuclear weapons, chemical weapons, cluster munitions — has been accompanied by a lag in international rules, followed by rebuilding. AI militarization is no exception, but its speed and opacity make the task of rebuilding rules all the more urgent.

---

One-sentence takeaway: When AI decides who to strike, the two core safeguards of international humanitarian law are breaking down — and no one knows who fills the accountability black hole.

Join the conversation: What role should AI play in military decision-making — "advisory support" or "autonomous decision-making"? Share your reasoning below.