TO HOME

What Are AI Lethal Autonomous Weapon Systems (LAWS)?

― What Are AI Lethal Autonomous Weapon Systems (LAWS)? The reality of killer robots that can decide targets without humans―Lethal Autonomous Weapon Systems (LAWS), also known as “killer robots,” are systems that process data collected through sensor inputs, then detect and apply force to a target based on that data. In other words, through artificial intelligence, they can select and attack targets on their own without direct human intervention.. . . . .Have you ever heard the term “killer robots”?Most people probably imagine a scene from the movie Terminator, robot armies moving without human commands, machines finding targets and attacking on their own.That is why, somewhere in our minds, we want to believe this is still “a future that has not yet arrived.”But what if that future has already arrived much sooner, and much closer, than we expected?Right now, a very different kind of war is already unfolding on the battlefields of Ukraine.Since Russia’s invasion in 2022, drones that were initially launched for reconnaissance have gradually turned into attack weapons carrying explosives. Soldiers now sit inside trenches, looking at screens on smartphones and tablets.When cameras attached to FPV (first-person view) drones capture the battlefield in real time and transmit the footage to a screen, soldiers watch the video and remotely pilot the drones toward their targets. No special equipment is needed commercial smartphones and apps are often enough.War has gradually begun to feel more like a game.But the more chilling change came next.Some drones have now started moving on their own.AI has begun analyzing targets by itself and even suggesting attack priorities.The scale is also expanding at a frightening speed.Ukraine has set a plan to purchase around 4.5 million FPV drones in 2025, while Russia deployed around 1,000 drones and strike assets in a single day in March 2026.Drones have now become wartime supplies consumed in mass quantities every day, almost like artillery shells.How Fast Has AI Drone Warfare Expanded?▲ These two figures are based on different units of measurement: one refers to an “annual procurement plan,” while the other refers to “daily deployment volume.”Ukrainian President Volodymyr Zelenskyy also directly addressed the dangers of AI weapons in a speech at the United Nations.He warned that “regulation is failing to keep up with the speed of technology,” emphasizing that the militarization of AI could become a new international security crisis.“Killer robots are not a future technology. They are a technology that has already begun entering the battlefield.” ▲ Source: Campaign to Stop Killer Robots― What Is the Crucial Difference Between Drones and AI Weapons? The definition of Lethal Autonomous Weapon Systems (LAWS)―In fact, many people tend to think drones and AI weapons are more or less the same.But there is one crucial difference hidden between the two.Who makes the final decision?Ordinary drones are controlled by humans. No matter how far away the operator may be, it is ultimately a human being who presses the final button.But Lethal Autonomous Weapon Systems (LAWS) are different.These systems can find targets, analyze them, and even carry out attacks on their own.Ordinary Drones vs. Lethal Autonomous Weapon Systems (LAWS): Who Makes the Decision? Ordinary Drones Human-controlledHuman verifies the targetA human makes the final attack decision Lethal Autonomous Weapon Systems (LAWS) AI detects and analyzes targetsAI suggests attack prioritiesHumans only approve The International Committee of the Red Cross (ICRC) defines autonomous weapons this way: “Autonomous weapon systems select and apply force to targets without human intervention.”— International Committee of the Red Cross (ICRC) Did you notice the most chilling part of that sentence?It is the phrase “without human intervention.”That is why the United Nations and the ICRC continue to emphasize one key concept: Meaningful Human Control. ① Humans must understand the context of an attack.② Humans must make the final decision.③ Humans must be able to intervene or stop the system at any time.④ If something goes wrong, humans and states must be held responsible.But here is the frightening fact.Once AI weapons emerge, they can spread extremely quickly.Unlike nuclear weapons, they do not require massive facilities. They are far cheaper and much easier to access.― A Machine Decides “Who Will Die” The cases of Lavender in Gaza and Palantir in Ukraine―You have probably heard of Professor Geoffrey Hinton, often called the “Godfather of AI.”In his speech at the 2024 Nobel Banquet, he issued this warning: “In the near future, AI could be used to create terrible new weapons that decide by themselves who to kill.”— Geoffrey Hinton, 2024 Nobel Banquet Speech Until now, war has at least had one final safeguard: human judgment.Soldiers were required to distinguish whether a target was a civilian or not. And if a wrongful attack occurred, there was someone who had to take responsibility.But now, AI is beginning to recommend targets, and humans are beginning to approve those judgments.Two battlefields show this reality clearly.First, let’s look at the controversy over “Lavender” in Gaza.In 2024, +972 Magazine reported that the Israeli military had used an AI system called Lavender to classify around 30,000 people as potential targets.Human review was reportedly limited, and allegations emerged that lists generated by AI had become deeply involved in the actual process of deciding attacks. If AI classified tens of thousands of people as dangerous individuals, and humans simply approved those results at speed, can we really say that “humans made the judgment”?And if a mistaken strike occurs, who is responsible?Algorithms do not take responsibility.The battlefield in Ukraine has gone one step further. Palantir’s platform, Gotham, integrates drone footage, satellite images, and battlefield data into a single system.AI then analyzes this information to identify the locations of Russian tanks, artillery positions, ammunition depots, and command posts. It even suggests which targets should be attacked first to achieve the greatest strategic effect.Commanders make decisions while looking at a screen filled with coordinates.And with a single click, that decision can lead to a real explosion.In the end, the truly frightening part of AI weapons may lie somewhere else:a form of warfare in which humans gradually stop thinking for themselves.― AI Lethal Autonomous Weapons Are an International Security Issue Four risks the United Nations is concerned about―AI can process far more data than humans.It does not get tired, and it is not swayed by emotion.But is that necessarily a good thing?As AI increases the speed of war and reduces casualties among a country’s own soldiers, the threshold for choosing war may also become lower.We may be entering a world where attacks can be decided more easily, with less burden.This is exactly why the United Nations views AI lethal autonomous weapons not simply as a technological issue, but as a matter of international security.Four Major Risks of Autonomous Weapons Warned by the United Nations Unclear accountability when mistaken strikes occur Reduced soldier casualties → lower threshold for war Hacking and data errors leading to deadly attacks Spread of cheap autonomous drones to terrorist organizations ▲ Based on the 2024 UN Secretary-General’s report 「Lethal Autonomous Weapons Systems」― Why Haven’t They Been Banned Yet? The security dilemma and the dual-use problem―If these weapons are so dangerous, wouldn’t it be enough to simply ban them?In fact, organizations such as the United Nations, the International Committee of the Red Cross (ICRC), and Human Rights Watch have consistently called for strong international norms.But reality is far more complicated than it may seem.The biggest reason is one very old emotion:distrust.States quietly ask themselves the same question:“What if we stop first, but our opponent keeps developing them?”Major powers such as the United States, China, and Russia see AI as a core element of future military power.In the end, everyone knows these weapons are dangerous.And yet, no one wants to let go first.In international politics, this situation is called the security dilemma.The bigger problem is that AI weapons are far more difficult to control than nuclear weapons.Nuclear weapons leave physical traces, such as uranium enrichment facilities. But with AI weapons, the boundary between civilian technology and military technology is extremely blurry.A technology developed as “drone AI for disaster rescue” can quickly turn into an “automatic target-tracking weapon” somewhere else.This is called the dual-use problem.― The UN’s Warning: This Is a Matter of Human Survival The 2024 UN Secretary-General’s report and António Guterres’ remarks―▲ This image is unrelated to any real person.The United Nations does not view AI lethal autonomous weapons simply as a matter of new military technology.It treats them as an issue of international security and human dignity.The 2024 UN Secretary-General’s report, Lethal Autonomous Weapons Systems, is an official document that brings together the views of governments, international organizations, the ICRC, civil society, the scientific community, and industry.The report emphasizes three core points. First, international law must be upheld even in war. Wars must distinguish between civilians and combatants, and excessive attacks must be avoided. But the question remains: can AI properly make these kinds of contextual judgments?Second, human responsibility must be preserved. When a wrongful attack occurs, the explanation “the algorithm made that decision” cannot replace accountability.Third, new international norms are needed. Killing without human control could shake the very foundations of the international order.  UN Secretary-General António Guterres has described this issue in firm terms:“Politically unacceptable and morally repugnant.”