A defense technology company operating in Ukraine is demonstrating how artificial intelligence (AI) is reshaping modern warfare through autonomous drone systems. Norda Dynamics, a Ukrainian-Estonian firm, has developed Underdog, a terminal-guidance system that enables drones to autonomously complete the final phase of a strike mission. According to CEO Nazar Bigun, this system is widely deployed in combat across Ukraine and achieves a success rate exceeding 90% once operators hand control over to the AI.

Ukraine has emerged as a testing ground for AI applications in military operations, harnessing the technology both for data analysis and weapon automation. The scale of data collected on the battlefield is immense; a single Ukrainian brigade generates the equivalent of 10 months’ worth of video daily, while the country’s battlefield-management system, Delta, processes quadrillions of bytes every day. Advanced military units use proprietary AI systems to transform this information into actionable intelligence.

Since March, with financial support from the United Kingdom, Ukraine has established A1, an AI integration center within its Defense Ministry, led by CEO Danylo Tsvok. The center has rolled out Spectr, an AI platform that can swiftly interpret battlefield data and undertake tasks traditionally performed by multiple personnel. It also empowers tactical units to access sensitive data offline. Upcoming features include operational planning tools offering commanders a range of options with estimated success probabilities and predictive capabilities to forecast enemy movements.

While these advancements offer tactical advantages, Ukrainian military personnel maintain a cautious stance toward AI. Commanders retain ultimate responsibility for decision-making, with AI systems designed to explain their reasoning and highlight data limitations. Yaroslav Azhnyuk of Kyiv-based companies The Fourth Law and Odd Systems, which develop autonomous drone and interceptor technologies, emphasized that fully autonomous weapons involve replacing human roles in the decision loop but stressed that mission parameters remain set by humans.

The use of autonomy in weaponry exists on a continuum. Systems like Underdog rely on human-identified targets and are evaluated as potentially more precise and less prone to collateral damage compared with traditional artillery. More advanced autonomy, projected in the near term, could see drones autonomously conducting entire strike missions—from takeoff to target identification to weapon release—within human-established boundaries.

Autonomy is particularly entrenched in Ukraine’s air defense, where the absence of commercial flights reduces civilian risks. Commanders can define operational zones in which AI-enabled systems engage hostile targets autonomously, a concept poised to extend to frontline ground operations under carefully controlled parameters.

These developments are transforming the battlefield landscape in Ukraine, shifting it from a static front line to a dynamic "kill zone" where movement risks immediate targeting. Autonomous drones are likely to expand this zone’s reach, while AI-driven analysis accelerates strike decisions and adaptation. The integration of AI and autonomous weapons delivers a strategic edge that can compensate for limited manpower or magnify existing advantages.

Though concerns persist among some Western technologists about autonomous systems acting unpredictably, Ukrainian military AI developers express deeper apprehension about the risk of authoritarian adversaries gaining the upper hand in the AI arms race. As Tsvok put it, “We can’t avoid using AI... The highest risk in Ukraine now is, don’t take the risk.”