Risk-Based Approach to AI Implementation in Ukraine's Defense Sector: Conceptual Framework

The article discusses the conceptual foundations of a risk-oriented approach to introducing artificial intelligence (AI) into Ukraine's defense sector. The study's relevance stems from the rapid development of AI technologies in military affairs and the need to counter the prevailing enemy. Thus, the article aims to identify and substantiate the fundamental principles of a risk-oriented approach to integrating AI into Ukraine's military. The primary method used in this study is analysis, which allows to consider non-technical internal risks, which are less obvious and potentially more dangerous than technical and external risks. The article identifies and categorizes key non-technical risks arising from the introduction of AI in the military. Specifically, it identifies the risks associated with underestimating threats, the uncertainty of AI development, the negative impact on humans (e.g., reduced cognitive skills and excessive trust), and problems related to automatic translation and decision-making. The article also considers external risks arising from the enemy's use of AI, such as increased intelligence effectiveness, cyberwarfare, propaganda, and data poisoning. The authors substantiate the necessity of creating a register and map of these risks for further management. The article's materials are practically valuable for developing an AI policy for the military, creating a risk register and map, and for further work with them.