This research presents an AI security assessment model, developed as a computational ontology grounded in Knowledge Representation and Reasoning (KRR) principles and constructed following a Design Science Research (DSR) methodology. The model represents the complex landscape of AI systems used in the military domains, encompassing systems such as AI-enabled weapon systems and AI-based decision support systems. At the same time, it incorporates important concepts like threats, attack vectors, vulnerabilities, risks, impacts, and military-specific operational requirements. To this end, the model provides a formalized, machine-interpretable structure that enables advanced reasoning over security relationships and supports automated risk assessment and decision-making processes. To demonstrate the model{\textquoteright}s effectiveness, the model is instantiated in a representative military use case, illustrating its capacity to identify and assess security risks, support compliance with operational and legal requirements, and inform the development of robust countermeasures. The results show the importance of building such efforts that assure both knowledge sharing and reasoning as well as decision-making support and provide simulation capabilities in AI-driven military environments.
Maathuis, C & Cools, K 2025, Military AI Security Assessment Model. in 2025 Cyber Awareness and Research Symposium (CARS). 2025 Cyber Awareness and Research Symposium, CARS 2025, IEEE, Grand Forks, pp. 1-7. https://doi.org/10.1109/CARS67163.2025.11337714
Maathuis, C., & Cools, K. (2025). Military AI Security Assessment Model. In 2025 Cyber Awareness and Research Symposium (CARS) (pp. 1-7). (2025 Cyber Awareness and Research Symposium, CARS 2025). IEEE. https://doi.org/10.1109/CARS67163.2025.11337714
@inproceedings{5c0ba133a1cd4895a0a7b73cc5ef4acd,
title = "Military AI Security Assessment Model",
abstract = "This research presents an AI security assessment model, developed as a computational ontology grounded in Knowledge Representation and Reasoning (KRR) principles and constructed following a Design Science Research (DSR) methodology. The model represents the complex landscape of AI systems used in the military domains, encompassing systems such as AI-enabled weapon systems and AI-based decision support systems. At the same time, it incorporates important concepts like threats, attack vectors, vulnerabilities, risks, impacts, and military-specific operational requirements. To this end, the model provides a formalized, machine-interpretable structure that enables advanced reasoning over security relationships and supports automated risk assessment and decision-making processes. To demonstrate the model{\textquoteright}s effectiveness, the model is instantiated in a representative military use case, illustrating its capacity to identify and assess security risks, support compliance with operational and legal requirements, and inform the development of robust countermeasures. The results show the importance of building such efforts that assure both knowledge sharing and reasoning as well as decision-making support and provide simulation capabilities in AI-driven military environments. ",
author = "Clara Maathuis and Kasper Cools",
note = "Publisher Copyright: {\textcopyright} 2025 IEEE.",
year = "2025",
doi = "10.1109/CARS67163.2025.11337714",
language = "English",
isbn = "979-8-3315-9629-3",
series = "2025 Cyber Awareness and Research Symposium, CARS 2025",
publisher = "IEEE",
pages = "1--7",
booktitle = "2025 Cyber Awareness and Research Symposium (CARS)",
}