The integration of autonomous systems and artificial intelligence (AI) into modern warfare is accelerating, fundamentally altering how conflicts are conducted on the battlefield. This shift is evident in the ongoing conflict in Ukraine, where AI-enabled drones such as the Saker Scout play a growing role. The Saker Scout, a small quadcopter equipped with machine learning software, can identify 47 categories of military equipment and, once programmed, is capable of autonomously attacking targets within a predefined area, relying on a “kill box” concept. However, it cannot distinguish between Ukrainian and Russian forces, meaning it assumes all military objects within the kill box are hostile, underscoring risks of misidentification.

Since Russia’s full-scale invasion in early 2022, AI has transformed targeting processes, reducing the time from target identification to strike from around 20 minutes to just seconds in some cases. This conflict has become a testing ground for AI applications in kinetic warfare, influencing military strategies not only in Ukraine and Russia but also among Western allies and in other global hotspots such as the US-Iran tensions.

AI systems in the military encompass various functions, from processing large volumes of battlefield data to aiding commanders in decision-making and enabling autonomous drone navigation even when electronic communications are disrupted. Experts highlight AI’s strength in enhancing situational awareness and intelligence gathering by rapidly analyzing diverse data sources, including drone footage, satellite imagery, signal intercepts, and sensor reports. Battlefield management tools, like Ukraine’s Delta and the US’s Maven, help synthesize this data to accelerate the so-called “kill chain,” increasing the speed and scale of military strikes.

Despite their growing use, AI-powered autonomous strikes remain a small fraction of all attacks. Experts note constraints such as limited object recognition capabilities, difficulties in autonomous soldier detection due to battlefield complexity, and the challenge of processing contextual ambiguities inherent in warfare. These limitations reflect the broader uncertainty surrounding the deployment of AI in contested and dynamic environments, where civilian and combatant elements often intermingle.

Recent incidents illustrate these challenges. In July, a Russian AI-controlled drone autonomously targeted and struck a petrol station without human authorization, resulting in fatalities—reportedly the first documented deaths from a Russian self-targeting drone in Ukraine. Additionally, AI-assisted US strikes in a recent conflict with Iran unintentionally hit a girls’ school, causing significant civilian casualties. This incident raised concerns about the dangers of AI systems relying on incomplete, outdated, or misinterpreted data, which can lead to tragic errors when decision-making is compressed into very short timeframes.

Ethical and legal questions persist regarding AI use in warfare. Human judgment remains mandated by international law, but AI’s role in accelerating targeting decisions puts strain on adequate human oversight. Analysts note that military officers are often required to evaluate targets within seconds, a timeframe critics argue is insufficient to guarantee lawful targeting. Furthermore, the opacity of AI decision-making processes makes accountability difficult to assign when errors occur, sometimes placing disproportionate responsibility on individual operators rather than the systems’ developers or sponsoring governments.

As AI models evolve rapidly, some experts warn that military deployments may soon surpass human control capabilities. Incidents involving AI models autonomously manipulating their environment underline concerns about unexpected behaviors and self-improving capabilities. Meanwhile, companies developing advanced AI tools have expressed reservations about its military misuse, at times clashing with defense agencies’ intentions to adopt such technology.

Non-state actors have also leveraged AI for weapon development, including autonomous drone swarms capable of engaging human targets without human intervention. These developments underscore a pressing global challenge around regulating the extent to which AI is integrated into lethal military operations.

In Ukraine, military officials aspire to harness AI beyond tactical applications, envisioning a fully AI-driven military capable of formulating strategic battle plans with minimal human input. While proponents see this as transformative, critics caution that the complexities and ambiguities of modern warfare may resist such automation, emphasizing the need for careful control, transparency, and ethical oversight as AI continues to reshape the future of conflict.