Usage of artificial intelligence in the aviation of the Armed Forces of Ukraine

Abstract

The paper systematizes current directions of implementing artificial intelligence (AI) in the aviation of the Armed Forces of Ukraine. It covers architectures for autonomous control, real-time ISR processing, predictive maintenance, adaptive electronic warfare (EW), and decision support systems. National and international initiatives are reviewed (Delta, SkyAssist, SHARK, Project Maven, ARTEMIS, NATO), demonstrating convergence between military and dual-use AI technologies. Recommendations are provided for stepwise integration, unified data standards, secure communications, and cyber-resilience. The research contextualizes technological advances within the operational realities of contested environments, including bandwidth-limited data links, degraded GNSS, dynamic electromagnetic interference, and constrained onboard computing. Proposed approaches emphasize layered modular architectures that combine edge inference with distributed fusion nodes, hybrid assurance frameworks integrating data-driven models with formal safety envelopes, redundancy logic, and human-on-the-loop supervision. Discussed methods include efficient perception networks, model compression and quantization for edge deployment, reinforcement learning for adaptive control, sim-to-real transfer for unmanned operations, and physics-informed prognostics for predictive maintenance. Legal, ethical, and certification aspects are analyzed with focus on explainability, trustworthy datasets, V&V practices, and protected model provenance. Finally, a phased roadmap outlines incremental trials, shared data registries, digital testbeds, simulation environments, coalition-aligned validation, and coordinated operator training as key enablers for scalable AI adoption across Ukraine’s defense aviation ecosystem. The study highlights governance structures, interagency cooperation, and knowledge transfer mechanisms as essential for sustainable AI-driven transformation of command, control, and aviation support domains