The use of Autonomous Weapon Systems has changed the dynamics of warfare in recent times. The rise of Artificial Intelligence has reshaped military power, strategies and decision-making in war. AI is transforming modern combat through greater speed, precision and low-cost deployment.

Autonomous weapons systems (AWS) are changing the nature of modern warfare through AI-enabled decision-making. As per the United States Department of Defense Directive 3000.09, autonomous weapons systems are platforms capable of ‘selecting and engaging targets without further intervention by a human operator’, a definition that anchors the concept of autonomy specifically to the targeting functions.[1] AWS combines artificial intelligence with lethal platforms like loitering munitions, drones, AI-powered surveillance systems, and swarm drones that have capabilities to identify and strike targets with limited human control. The conflicts in Ukraine and Gaza have become testing grounds for AWS, showing how these technologies operate in both high-intensity and asymmetric conflicts.

Autonomous weapons systems are mainly considered in two key respects: how much control humans have over decisions to use lethal force, and how much autonomy the system itself has. AWS are commonly grouped into three categories: ‘human-in-the-loop’, ‘human-on-the-loop’ and ‘human-out-of-the-loop, ‘ depending on the degree of human involvement. Many autonomous systems currently in use fall into the ‘human-on-the-loop’ category; even though they can carry out pre-planned actions involving lethal force, human control can intervene if needed.[2] This classification raises significant International Humanitarian Law concerns regarding accountability, proportionality and compliance.

AWS as Force Multipliers

A Force multiplier is a term used in military doctrine that refers to the capacity of a technology, technique, or organisational arrangement to expand and amplify the combat effectiveness of a given force well beyond its numerical strength alone.[3] The force-multiplication effect of AWS operates across four distinct and interconnected dimensions: speed, persistence, precision and expendability. Speed refers to the capacity of autonomous weapons to execute targeting cycles at machine speed, overcoming the physiological and cognitive delays in human decision-making on the battlefield. Loitering munitions with AI guidance can detect and engage a target in a much shorter time than those controlled by humans. AWS acts faster because it doesn’t get tired or confused as humans do.[4]

Persistence is the second dimension. As humans need rest, resupply, and rotation, autonomous platforms can conduct continuous surveillance of targets for longer durations, in hours or days. A persistent autonomous intelligence, surveillance and reconnaissance (ISR) system can reduce intelligence collection gaps. As observed in Ukraine, persistent drone surveillance has made it really hard for either side to move a group of people on the ground without being seen. Similarly in Gaza, persistent aerial surveillance enabled Israel Defense Forces to keep an eye on designated targets across urban areas, creating a challenge for the Gaza military. This has changed the way wars are traditionally fought.[5]

Precision and expendability are the third and fourth dimensions, respectively. Compared to traditional guided munitions, AI-guided loitering munitions deliver superior precision at a significantly lower cost.[6] Expendability changes the cost equation for high-risk missions. An expendable autonomous platform can target highly defended areas without the fear of losing a pilot.[7] Collectively, these four aspects illustrate the importance of AWS as a force multiplier.

Ukraine: Attrition Warfare and Autonomous Scale

The Russia–Ukraine conflict has emerged as one of the prime examples for testing autonomous and semi-autonomous systems in a conventional war against a peer adversary since World War II. Ukraine’s early deployment of the Turkish-made Bayraktar TB2 Unmanned Aerial Vehicle (UAV) demonstrated that AI-enabled unmanned systems, when combined with real-time ISR and precise targeting data, could effectively weaken the enemy’s conventional armoured units. In the first week of the full-scale invasion, TB2 operations struck Russian Buk surface-to-air missile systems, armoured vehicles, and other military targets, highlighting the operational value of integrating AWSs with real-time battlefield intelligence.[8]

As the war escalated between Russia and Ukraine, both sides started making adaptive changes to the situation. Russia swiftly acquired and deployed the Iranian-made Shahed-136 loitering munitions to target Ukraine’s critical infrastructure because of their low cost and high destructive potential. Each of these drones costs them approximately US$ 20,000 to US$ 50,000 per unit.[9] Russia has been deploying these drones to overwhelm Ukrainian air defences, and later this number rose approximately from 80–100 to 100–200.[10] To overcome this challenge, Ukraine developed a domestic manufacturing unit of substantial scale during wartime. Ukraine planned to produce around 1 million first-person view (FPV) drones in 2024. These FPV drones became an important weapon against Russian tanks and soldiers on the front itself.[11]

Ukraine has also developed and deployed AI-enabled command-and-targeting software to reduce sensor-to-shooter time. This includes the Delta battlefield management system, developed by Ukraine and designed to operate in accordance with NATO interoperability standards. It uses multiple sources of information, such as satellite images, signals........

© IDSA