Artificial Intelligence and Modelling & Simulation in the military affairs: From Lifecycle Transformation to Decision Support
The rapid evolution of Artificial Intelligence (AI), alongside increasing operational complexity and the growing demand for timely, data-driven decision support, is reshaping the role of Modelling and Simulation (M&S) within NATO. While AI and M&S are increasingly developed and applied as complementary technologies, their integration across the M&S lifecycle remains fragmented, with significant challenges concerning model development, data generation, scenario construction, orchestration, verification, validation and accreditation (VV&A), explainability, and human oversight. This paper examines how AI can be integrated as a cross-cutting capability throughout the M&S lifecycle, with the aim of strengthening rather than replacing established modelling and simulation practices. The work addresses the challenge of connecting heterogeneous data, models, analytical methods and AI capabilities while preserving the credibility, traceability and reproducibility required for defence applications. Methodologically, the paper develops a conceptual framework by mapping AI-enabled capabilities onto the principal stages of the M&S lifecycle, from operational problem definition and requirements analysis through conceptual modelling, implementation, execution, VV&A and model evolution. The framework is complemented by an agentic orchestration architecture in which specialised AI capabilities interact with data sources, simulation models, analytical tools and human experts, enabling more flexible coordination of complex M&S activities. The paper further examines representative applications, including synthetic data generation, scenario generation, intelligent agents, multi-domain operations and wargaming, and considers the assurance requirements associated with AI-supported M&S. The resulting framework positions AI and M&S as mutually reinforcing capabilities within an integrated ecosystem rather than as independent technological domains. The proposed conceptual framework provides a basis for structuring future NATO AI–M&S architectures and identifying the technical, organisational and assurance mechanisms required for their adoption. Its principal implication is that successful integration will depend not only on the performance of individual AI techniques, but also on the ability to establish trustworthy, traceable and human-supervised processes across the complete M&S lifecycle.