By Shahzad Ali
Over a disputed boundary, a drone scans for a person carrying something that could be a weapon. No soldier decides whether to fire. No one weighs the risk to nearby civilians. Only code runs — trained on data, executing a decision loop it was built for. The strike happens in milliseconds. The question of lawfulness comes later, if it comes at all.
This isn’t a scenario from a science fiction script. It’s the direction warfare is already moving, and international humanitarian law (IHL) is not moving fast enough to meet it.
Law Built for Human Judgment
IHL was never built for machines. Its core principles — distinction, proportionality, precaution — all assume a human capable of moral judgment under pressure. Distinction means telling combatant from civilian in real, often unclear conditions.Proportionality involves considering the expected military advantage and the likely civilian damage, and this is not made purely from data: it is dependent on context and purpose. Precaution demands situational awareness no sensor can fully replace.
Autonomous Weapons Systems (AWS) don’t perform these as judgments — they perform them as calculations, shaped by whatever parameters their designers built in. A system can be technically precise, hitting exactly what it’s told to hit, while remaining incapable of the human moral reasoning IHL requires.
This Is Already Happening
This isn’t a distant frontier. AI-driven targeting systems have reportedly helped generate and prioritize strike targets in recent conflicts, compressing decisions that once took human analysts far longer.Weapons that can detect and attack targets with little or no human intervention are already used in the conflict zones. But adoption is outrunning regulation, and the difference between what militaries are deploying and what the law clearly allows to get bigger each year.
The Accountability Vacuum
Perhaps the greater danger isn’t the targeting error itself, but what happens after. When a human soldier kills unlawfully, a chain of responsibility exists, at least in principle: person, commander, court. When an autonomous system selects and strikes a civilian target, that chain breaks. Is the programmer responsible for a flawed model? The commander who deployed a system they may not fully understand? The state that authorized it? IHL’s enforcement architecture assumes an identifiable human decision-maker. Autonomous weapons risk a permanent accountability vacuum — devastating harm with no one clearly answerable.
The Law Is Standing Still
The international community has not been silent, but it’s been slow. The UN’s Group of Governmental Experts on Lethal Autonomous Weapons Systems have been meeting for more than 10 years, but have yet to come up with a treaty. Civil society organisations have advocated for meaningful human control over the use of force, including the Campaign to Stop Killer Robots.. But major military powers have resisted firm commitments, wary of limiting an edge before rivals do the same. The result is a legal vacuum filled not by careful multilateral deliberation, but by whatever states choose to deploy.
What the Law Must Require
The answer isn’t to ban all military AI — some applications genuinely improve precision and reduce civilian risk. The answer is a binding international requirement of meaningful human control over any decision to use lethal force: a human must be able to understand, verify, and override a targeting decision before it executes, not simply supervise a system that has already decided. It also means establishing clear legal accountability before deployment, not after the first unlawful strike forces the question.
IHL has always adapted, gradually, to new methods of warfare — chemical weapons, landmines, cyber operations. Autonomous weapons deserve no exception, and need one more urgently than any before, because the harm they cause cannot be undone. Every year without a binding framework is another year in which life-and-death decisions in armed conflict are made by algorithms accountable to no one, not soldiers accountable to law.
The technology is not waiting for the law to catch up. Neither should the international community.
Also Read: A New Digital Horizon — Google’s Arrival and the Promise for Pakistan’s Youth
Today's E-Paper