Conflicts persist, sustained by false assumptions and manufactured narratives.
Across every military domain, artificial intelligence is quietly reshaping the nature of war — not merely by making weapons faster, but by turning information itself into a weapon. Governments have been embedding machine learning into autonomous systems for decades, yet no binding international framework exists to govern where these technologies lead. The question humanity now faces is not whether AI will fight its wars, but whether wisdom can outpace the algorithms before the consequences become irreversible.
- AI-powered weapons systems now operate across air, land, and sea with a speed and autonomy that increasingly sidelines human judgment in life-and-death decisions.
- Cognitive warfare — deepfakes, manufactured narratives, algorithmically amplified disinformation — is actively distorting reality for civilians and prolonging conflicts built on falsehoods no human authored.
- Lethal Autonomous Weapons Systems carry compounding dangers: vulnerability to cyberattack, catastrophic infrastructure dependence, and the concentration of lethal power in whoever controls the code.
- Civil society, researchers, and think tanks are sounding alarms, but governments continue to classify most AI weapons as 'defensive,' obscuring the blurring line between protection and aggression.
- The window for establishing global governance frameworks is narrowing fast, and no binding international agreement yet exists to prevent these systems from reaching hostile actors or escaping meaningful human oversight.
Militaries worldwide are racing to embed artificial intelligence into their weapons systems, drawn by a straightforward promise: speed and efficiency. AI-powered defense networks process sensor data in real time, detect threats with greater accuracy than human operators, and direct autonomous drones and unmanned vehicles through combat operations with minimal human involvement. Machine learning algorithms, deep neural networks, and transformer architectures together give weapons something that resembles cognition — the ability to perceive, decide, and act faster than any human could.
But speed carries a shadow. The same systems enhancing situational awareness are now central to cognitive warfare — the weaponization of information itself. During the US-Iran conflict, both sides deployed AI to generate deepfakes and flood social media with content designed to inflame divisions and celebrate manufactured victories. The goal was not always to kill, but to confuse. Civilians, submerged in AI-generated disinformation, lost the ability to understand what was actually happening. Conflicts persisted, sustained by false assumptions that no human mind created but that human minds could not easily distinguish from truth.
The risks compound further with Lethal Autonomous Weapons Systems, which lack human judgment, remain vulnerable to cyberattack, and concentrate power dangerously in the hands of those who control the algorithms. Nonviolence International Southeast Asia has documented how governments have been moving toward these systems since the Cold War, with the pace accelerating sharply in recent years. Military planners point to genuine benefits — realistic training simulations, active missile protection systems — but the disadvantages are growing harder to ignore as the technology spreads beyond controlled hands.
Right now, no binding international framework exists to govern autonomous weapons or AI-powered cognitive warfare. Governments claim most systems are defensive, but the line between defense and offense dissolves when the battlefield is the human mind. The window to establish global governance is closing, and prudence is no longer a luxury — it is a necessity.
Militaries around the world are racing to embed artificial intelligence into their weapons systems, drawn by a simple promise: speed and efficiency. An AI-powered defense network can process sensor data in real time, detect threats with greater accuracy than human operators, and direct autonomous drones, unmanned vehicles, and ground systems to conduct reconnaissance and combat operations with minimal human involvement. The technology works. It works so well that governments have been quietly experimenting with increasingly sophisticated AI functions in weapons for decades, accelerating sharply in recent years as the technology matured.
The mechanics are becoming familiar to defense planners. Machine learning algorithms—systems that learn patterns from data without explicit programming—power these autonomous platforms. Deep neural networks, inspired by the structure of the human brain, can recognize patterns and make decisions that approach human-level performance. Convolutional neural networks process visual data from cameras and sensors. Recurrent networks handle sequential information. Transformers, the newest architecture, process information in parallel, overcoming limitations that plagued earlier systems. Together, these tools give weapons systems something that looks like cognition: the ability to perceive, decide, and act faster than any human could.
But speed has a shadow. The same AI systems that enhance situational awareness and cyber defense are now central to what researchers call cognitive warfare—the weaponization of information itself. During the US-Iran conflict, both sides deployed AI to generate predictive analyses of enemy movements, to create deepfake images and video so convincing that they blur the line between documentation and fabrication, and to flood social media with content designed to celebrate military victories and inflame internal divisions. The goal is not always to kill; it is to confuse. Civilians, drowning in AI-generated disinformation, lose the ability to understand what is actually happening on the ground. Conflicts that might otherwise end instead persist, sustained by false assumptions and manufactured narratives that no human mind generated but that human minds cannot easily distinguish from truth.
The risks compound. Lethal Autonomous Weapons Systems—LAWS, in the jargon—lack human judgment and intuition. They are vulnerable to cyber attack. They create a dangerous dependence on infrastructure and technology that can fail. And they concentrate power in the hands of whoever controls the algorithms. A Philippines-based think tank, Nonviolence International Southeast Asia, has documented how governments have been moving toward these systems since the Cold War, but the pace has accelerated. The concern is no longer theoretical: it is urgent and global.
Military planners point to genuine advantages. AI can create realistic virtual training environments where soldiers practice against intelligent opponents, learning strategy and tactics in safety before facing real combat. Simulations allow defense personnel to test plans and evaluate responses without risk. Active protection systems use AI to intercept incoming missiles and projectiles before they strike. These are not trivial benefits.
Yet the disadvantages are equally real, and they are growing harder to ignore. The technology is spreading. Researchers and military contractors are collaborating on AI weapons systems. Civil society is watching, and rightly worried. The question is no longer whether AI will be used in weapons—it already is, across air, ground, and water domains. The question is whether the world will establish governance frameworks before these systems reach hostile actors, before the algorithms escape human control, before cognitive warfare becomes the default mode of conflict.
Right now, there is no adequate international agreement. No binding framework exists to prevent the proliferation of autonomous weapons or to establish rules of engagement for AI-powered cognitive warfare. Governments claim most of these systems are defensive—designed to protect against incoming threats. But the line between defense and offense blurs when the weapon is information, when the battlefield is the human mind, when the cost of a mistake is measured in civilian lives lost to confusion and conflict that never needed to happen. Prudence and vigilance are not luxuries. They are necessities. The window to establish global governance is closing.
Notable Quotes
Speed and efficiency are commonly identified as the primary benefits of equipping weapons systems with AI technology— Nonviolence International Southeast Asia report
The Hearth Conversation Another angle on the story
Why does speed matter so much in military AI? Isn't accuracy more important than how fast a decision gets made?
Speed and accuracy are linked in ways that matter on a battlefield. A system that detects a threat in milliseconds instead of seconds can intercept a missile before it reaches its target. But you're right to push back—the real problem is that speed is being used to justify removing humans from the decision entirely. Once you do that, you've lost something you can't get back.
What's the difference between using AI to defend against incoming attacks and using it to spread disinformation?
One is a shield. The other is a weapon aimed at the mind. When you use AI to generate deepfakes and flood social media with false narratives, you're not trying to destroy infrastructure—you're trying to destroy people's ability to know what's true. And civilians have no defense against that.
Is there a way to use these technologies safely, or is the risk inherent?
The risk isn't inherent to the technology itself. It's inherent to the lack of rules. Right now, there's no international agreement on how these systems should be used or who should be allowed to build them. That's the real problem. Without governance, we're just waiting for the worst version of this to happen.
You mentioned that conflicts persist because of AI-generated disinformation. Can you give me a concrete example of how that works?
During the US-Iran conflict, both sides used AI to create content that celebrated their military successes and inflamed internal divisions in the other country. The content spread across social media faster than any human could fact-check it. Civilians couldn't tell what was real. The conflict didn't end because the information ecosystem was so polluted that people couldn't agree on basic facts anymore.
So the real danger isn't the weapons themselves—it's losing control of the narrative?
It's both. The weapons are dangerous because they reduce human judgment. But the narrative warfare is dangerous because it makes it impossible for humans to make good judgments at all. You've removed humans from the loop twice over.