Victory and defeat on the battlefield were once decided by the number of troops, the power of weapons, and the strength of fighter jets and missiles. Now artificial intelligence (AI), autonomous drones, robotic combat systems and technology that lets machines identify and strike targets have been added to that reckoning. As a result, the nature of war is changing fast.
This technological shift in defence is being called a new transformation of warfare. An important part of it is lethal autonomous weapon systems (LAWS): weapons that, in certain situations, can identify, select and decide to attack a target without direct human instruction.
There is an important distinction here, however. Not every unmanned aircraft or autonomous drone is a killer robot. A drone may fly by itself, set its own route or identify targets; but if a human takes the final decision to attack, it is not a fully autonomous lethal weapon. It is with this distinction in mind that we should look at how ready each country is, and how far the use of autonomous weapons has gone on the modern battlefield.
United States: heavy investment in AI-driven warfare
The US military has long researched autonomous and semi-autonomous weapons. Its warships carry technology such as the Phalanx close-in weapon system, which can automatically detect incoming missiles and other threats and take defensive action.
On 25 January 2023 the US Department of Defense updated Directive 3000.09 on autonomy in weapon systems. It stresses maintaining appropriate human judgment over the use of force, ensuring safety and complying with the law of war. The policy does not, however, ban research and development of autonomous weapons.
A US Congressional Research Service report published in March 2026 says US policy does not prohibit the development or use of lethal autonomous weapons. The Department of Defense has not publicly confirmed whether the country has built or deployed such weapons.
China: progress in unmanned combat aircraft and drones
China has stepped up investment in military drones, unmanned combat aircraft and AI-driven warfare technology. Its GJ-11 stealth unmanned combat aircraft is a much-discussed example, built to evade enemy air defences to carry out surveillance and strikes.
An important goal of China’s military technology is to use drones for surveillance, intelligence gathering, target identification and coordinated operations. Public information on how far these technologies can carry out lethal strikes fully autonomously is limited, however. Technical capability and the actual deployment of fully autonomous weapons therefore have to be considered separately.
Israel: AI in target identification and military decisions
Israel’s military uses advanced technology for surveillance, intelligence analysis, missile defence and target identification. Its Harop loitering munition and the Iron Dome missile defence system are notable examples.
In the Gaza war, Israel’s AI-driven data analysis systems called Habsora (The Gospel) and Lavender have caused international controversy. These systems were reported to identify and list potential military targets. It would not be right, however, to call them fully autonomous killer weapons as such.
In an explanation published in June 2024, the Israel Defense Forces said these technologies assist intelligence analysis. According to the IDF, a human role is maintained in the military decision-making process where the technology is used.
Russia and Ukraine: technology tested in real combat
The Russia-Ukraine war has become a major testing ground for the military use of modern drones and autonomous technology. Both sides are using surveillance drones, FPV drones, loitering munitions and electronic warfare.
On the battlefield in Ukraine, AI-driven target identification, autonomous drone navigation and technology that keeps a drone heading towards its target even when GPS signals are jammed have become important. Russia is also using loitering munitions such as the Lancet and various unmanned aerial vehicles.
But a drone being able to fly by itself is not the same as it selecting and attacking a lethal target without human instruction. Although the use of AI on the battlefield is growing, how much human control there is over each strike depends on the specific weapon and operation.
Russia: moving towards autonomous drones
For Russia, an important study in 2026 brought new information to light. In a report published on 13 April 2026, the Center for Strategic and International Studies (CSIS), a Washington-based research institution, said Russia is adding AI to military drones to increase their capacity for autonomous decision-making.
CSIS examined the technical features of a drone called the V2U. Citing Ukrainian technical analysis, CSIS said evidence had been found of computing and AI capabilities in the drone that could help it fly when cut off from communications and select targets itself. The researchers’ assessment is that Russia has probably used a fully autonomous drone on the battlefield.
Ukraine: using AI while keeping human control
Ukraine is also using AI to make its drones more effective on the battlefield. A Reuters report on 29 November 2025 said Ukrainian drone pilots are using AI to help identify and track targets even amid electronic jamming. Some drones can keep flying and identify targets through their cameras even when contact with the operator is lost.
Citing Ukrainian military officials, the report said the final decision to attack in these systems is being kept in human hands. AI target identification can still make mistakes, however, and its effectiveness depends on the environment and conditions. In other words, the Russia-Ukraine war is now not only a contest of unmanned drones but also a race to combine drones with AI, autonomous navigation and target-tracking technology.
Turkey: at the centre of the autonomous weapons debate
The Turkish-made Kargu-2 drone is one of the leading examples in the international debate over autonomous weapons. A 2021 UN Security Council report on Libya’s civil war referred to weapons capable of attacking targets autonomously, and the Kargu-2 was discussed in it.
There is disagreement, however, over whether the Kargu-2 can decide to kill people fully autonomously. An analysis by the International Institute for Strategic Studies (IISS) says that, according to the manufacturer STM’s description, the drone has autonomy in navigation and target identification but a human is meant to have a role in the decision to attack. The UN report’s account and the manufacturer’s statement differ on this.
China and South Korea: two kinds of preparation
China is working on autonomous navigation for military drones, coordinated drone operations and unmanned combat aircraft technology. The US Department of Defense’s 2025 annual report on China says that various Chinese military drones under development are claimed to have AI-based navigation and the ability to coordinate with other drones. Many systems, however, still need targets set in advance by humans, remote control or significant human input.
South Korea’s Samsung SGR-A1 robotic sentry system, meanwhile, has been discussed for many years. Built for surveillance in the border area with North Korea, it combines cameras, sensors and a weapon. What matters, though, is that the final decision to use the weapon is kept in human hands. For this reason it would not be accurate to present it as an example of a fully autonomous killer robot.
How far has it reached the battlefield?
So far, the most visible progress in autonomous warfare has come in a few areas:
- Autonomous navigation: drones can fly along set routes without constant human control.
- Target identification: vehicles, military installations and other targets are being identified with cameras and AI.
- Countering electronic jamming: some drones can keep operating with their own sensors and software even when communications or GPS signals are jammed.
- Coordinated drone operations: technology for running several drones together in a single operation is improving.
- Automated defence: automated defence systems are being used to detect missiles, rockets and drones and respond quickly.
There is no evidence, however, that all of these capabilities have been combined in a single weapon. Technology has limits, especially in distinguishing between combatants and civilians on a complex battlefield and in holding off an attack after reading the situation.
What international law says
As autonomous weapons spread, international discussion of how to control them has intensified. Talks on lethal autonomous weapons have been under way since 2014 within the framework of the UN Convention on Certain Conventional Weapons (CCW).
According to the UN, the Group of Governmental Experts on lethal autonomous weapon systems (GGE on LAWS) agreed in 2026 on a working characterisation, or conceptual framework, for discussing such weapons. It focuses on weapon systems able to identify, select and attack targets without direct human intervention.
No separate international legal protocol for autonomous weapons has yet come into force, however, which means there is no universal ban on them. The CCW’s Seventh Review Conference will be held in Geneva from 16 to 20 November, offering a chance to discuss a new international treaty or legal framework to control autonomous weapons and to decide next steps.
Countries remain divided. A Reuters report on 4 September 2026 said the United States and Russia had objected to a legally binding treaty, while several countries including Brazil and Ireland favour binding international rules. India and Turkey favour voluntary guidelines and an emphasis on existing international humanitarian law.
Human rights organisations, including the International Committee of the Red Cross, say the use of lethal weapons without meaningful human control could create serious problems for international humanitarian law, the protection of civilians and accountability.
The main risks of autonomous warfare
A major concern for experts is that there is no guarantee AI decisions on the battlefield will always be accurate. An AI-driven weapon could mistakenly identify a civilian as a military target. And if communications are cut or the weapon is hit by a cyberattack, its behaviour could slip out of control.
There are several other major risks. Mistaken target identification could see innocent people attacked and civilians killed. Assigning responsibility will become complicated: if an autonomous weapon commits a war crime, who is accountable, the manufacturer, the programmer, the military commander or the weapon’s operator? An arms race could intensify: if one country builds autonomous weapons, others may race to acquire the same technology.
The risk of such weapons falling into the hands of terrorist groups will also grow, as armed groups could carry out attacks by misusing commercial drones and readily available technology. If a weapon’s software is hacked or its data corrupted, terrible accidents could follow, so concern over cybersecurity will rise. War will also speed up: machines can make decisions much faster than humans, which could shrink the chance to contain a conflict before it spreads. For these reasons, not only the technical capability of weapons but also their use, control, accountability and international law have become equally important.
So is the world heading for fully autonomous war?
An analysis of the current situation shows that the world has not yet entered an era of fully autonomous war. But the trend towards reducing reliance on humans in important parts of warfare is clear.
Countries such as the United States, China, Russia, Israel, Ukraine and Turkey are building and using AI-driven military technology to varying degrees. In some cases weapons can fly, search for and track targets by themselves, and in some cases research is also under way into the ability to attack without human approval.
The Russia-Ukraine war in particular has shown that on future battlefields powerful weapons alone will not be enough; rapid data analysis, automated surveillance and coordinated use of drones will also become important. For fully autonomous lethal weapons, however, technical capability, reliability and international legal obligations remain big questions.
Autonomous warfare is no longer just a vision of the future. Its technological foundations have been laid and various applications are already being seen on the battlefield. But whether the world has reached the point of removing humans entirely and handing the final decisions of war to machines is still a matter of debate.
বাংলায় মূল প্রতিবেদন পড়ুন · Read the original Bengali report