Introduction
International humanitarian law governs the conduct of armed conflict between states and armed groups. It rests on a simple premise: humans make every decision on the battlefield. Soldiers identify targets, assess risks, and pull triggers themselves. For decades, this assumption held true across nearly every armed conflict. Technology first began testing this assumption through remotely piloted drones. Operators controlled these drones from thousands of miles away. A human still made the final decision to strike. This arrangement already stretched IHL’s traditional framework in new directions.
Artificial intelligence now pushes this problem much further. Modern weapons systems can select and engage targets independently. No human operator makes the final targeting decision in real time. The machine identifies a target and fires without waiting for approval. This shift creates a serious legal problem. IHL’s core principles, distinction, proportionality, and precaution, assume human judgment at the moment of attack. Autonomous systems remove that human judgment entirely. The law must now answer an uncomfortable question: can rules built for human decision-makers still govern machines that decide on their own?
IHL Foundations at Stake
The foundations of modern war rest firmly on the 1949 Geneva Conventions and also the 1977 Additional Protocol I. These treaties protect people who are not fighting and limit the suffering of all soldiers. Three core principles guide every military action: the vital rules of distinction, proportionality, and also the duty of precaution. Distinction requires every soldier to separate military targets from innocent civilians and civilian buildings at all times.
Article 48 of Protocol I makes this duty the primary obligation for any person conducting a military operation.Proportionality bars any attack causing civilian harm excessive to the direct military gain. Article 51 of Protocol I specifically defines what constitutes an indiscriminate or disproportionate attack on a target. Finally, precaution requires commanders to take all feasible steps to verify their targets and minimize any potential harm. Article 57 outlines the specific steps a commander must take before they launch any strike against an enemy. Every technological advance mentioned in this piece tests one or more of these three fundamental pillars of law. We must determine if these rules can survive when a machine rather than a human interprets the facts. These principles form the moral compass of the battlefield and we cannot afford to lose them to automation.
Drone Warfare- Existing Practice and Legal Friction
Armed drones are remotely piloted by humans and they currently sit inside a very difficult legal grey zone. These machines are not autonomous but the operators are often geographically and psychologically removed from the actual fighting. Critics often talk about a “PlayStation mentality” where operators might treat real life like a harmless video game. This emotional distance can lead to decision fatigue and might weaken the operator’s focus on the duty of precaution. We see this debate play out in the United States’ programs in places like Yemen, Somalia, and Pakistan.
Recurring controversies over civilian deaths have led to many critical reports from United Nations experts on extrajudicial killings. The conflict in Ukraine has now become a massive, live testing ground for both surveillance and strike drones. This large-scale use has pushed the global legal debate far beyond the earlier focus on simple counterterrorism efforts. Modern drone warfare forces us to reconsider how we apply the duty of precaution in real-time operations. The physical safety of the pilot changes the risk calculus and may encourage more frequent use of force. We must carefully monitor how these remote systems affect the mental state and legal compliance of military personnel. The law requires a level of deliberate judgment that can become strained by the speed of drone strikes.
From Remote-Controlled to Autonomous
There is a very clear line between a drone with a pilot and a truly autonomous weapon system. A human pilot still follows the traditional rules of accountability but an autonomous system breaks that vital chain. The International Committee of the Red Cross defines these systems as weapons that can select and engage targets alone. Once activated, these machines trigger a strike based on sensor data rather than a specific, human-identified target.
This is the heart of the legal problem because machines rely on a generalized “target profile” to fire. This approach sits very uneasily against the rule of distinction which requires a specific determination of a lawful target. A machine cannot feel the weight of a life-and-death decision at the exact moment an attack begins. The rupture occurs when we transfer the authority to kill from a reasoning human to a mathematical algorithm. Algorithms often struggle with the messy and unpredictable reality of a dynamic and changing urban battlefield. We cannot simply assume that a machine will correctly identify a surrendering soldier or a wounded person. This shift represents a fundamental change in how we understand the basic legality of a military strike. The lack of a human “in the loop” creates a risk that current laws simply cannot manage.
Can AWS Ever Comply With IHL?
The global community is split on whether these autonomous systems can ever truly comply with international humanitarian law. The Red Cross and a majority of countries argue that current rules are simply not clear enough today. They believe we need specific new rules or even total bans on certain types of lethal autonomous machines. On the other side, countries like the United States believe that the existing legal system is already robust. They argue that we do not need new binding restrictions to manage these emerging technologies on the battlefield.
However, a massive majority of states at the United Nations General Assembly recently voted for further global regulation. Only three countries opposed a 2024 resolution that called for more oversight of these increasingly complex weapons. United Nations Secretary-General António Guterres has called for a new legally binding treaty to be finished by 2026. He argues that we must prohibit weapons that operate without any human control or meaningful oversight at all. Right now, international groups are meeting to negotiate rules on predictability, traceability, and the need for human control. They want to ensure that if a machine is used, its actions can be explained and traced back. This ongoing negotiation shows that the world is still trying to find a common language for robotic warfare.
The Accountability Gap
The most difficult legal challenge is the accountability gap that appears when a machine makes a sudden mistake. If an autonomous system kills civilians in violation of the law, we must ask who bears the blame. Is it the commander who turned it on, the programmer who wrote the code, or the manufacturer? Current law assumes there is a human decision-maker whose intent and knowledge can be judged in a court. The Rome Statute focuses on the individual mental state of a person when they commit a war crime. Autonomous target selection disrupts this chain of command because a human did not pick that specific person. This means a war crime could happen without any person ever making a specific, illegal judgment during the strike.
We currently lack a clear way to punish a piece of software or a machine for its actions. Command responsibility usually requires the leader to have known about the crime and failed to stop it. If the machine acts in an unpredictable way, the commander might claim they had no way of knowing. This creates a situation where serious violations of international law could occur with no one being held responsible. We must close this gap before autonomous weapons become a standard part of every modern military arsenal.
Critical Analysis
We are not just facing a technical problem but a deep mismatch between law and our new reality. International humanitarian law is built for humans but the battlefield is increasingly populated by machines that make decisions. I believe that simply interpreting old laws is no longer enough to protect civilians in this digital age. The accountability gap is too wide and we truly need new binding laws to ensure that humans stay responsible. The biggest obstacle remains the fact that the most powerful nations are often the ones resisting these restrictions. This means that our global consensus might lag far behind the technology that is already deployed in conflicts.
The anthropocentric design of our current laws assumes that a human heart and mind are always present. When we remove the human, we remove the very foundation on which our entire legal system is built. We must prioritize human control as a legal requirement rather than just a technical preference in future treaties. Only a clear and binding international agreement can prevent the dehumanization of the battlefield in the coming years. The path forward requires a brave commitment to keeping human judgment at the center of lethal force. We cannot allow the convenience of technology to override our shared commitment to the laws of war.
Conclusion
The 2026 deadline for a treaty makes this a live and very urgent question for every nation today. These negotiations will determine if the core principles of humanity can survive in a world of robotic war. The law is currently racing to catch up to a reality that the original drafters could never imagine. If we fail to act, we risk losing the very rules that keep war from becoming total chaos. This is more than a legal debate; it is a fight for the future of our shared human values. We must ensure that the machines we build remain under the control of the laws we have written.
The outcome of the ongoing CCW meetings will shape the safety of civilians for many decades to come. Our goal must be to preserve the humanity that international law was meant to protect since the beginning. Let us move forward with a focus on accountability, transparency, and the absolute necessity of human intervention. The survival of the Geneva Conventions depends on our ability to adapt to this challenging new technological landscape. We must act now to ensure that the battlefield of the future remains governed by the rule of law.
References
- Geneva Conventions (1949).
- Additional Protocol I to the Geneva Conventions (1977).
- Convention on Certain Conventional Weapons (CCW) (1980).
- Rome Statute of the International Criminal Court (1998).
- UN General Assembly Resolution on Lethal Autonomous Weapons Systems (December 2024).
- António Guterres, “A New Agenda for Peace” (2023).
- CCW Group of Governmental Experts on LAWS, Sessions through 2026.
