A Systematic Evaluation of Artificial Intelligence's Impact on the Landscape of Threat Modeling

Abstract

This article systematically evaluates the influence of Artificial Intelligence (AI) on the landscape of Threat modeling in cybersecurity. The study involves an extensive review of relevant journal articles, books, and conference papers to comprehensively assess the current state of the field. By synthesizing existing literature, we identify and analyze the ways in which AI technologies are applied in Threat modeling methodologies. The evaluation explores the strengths and limitations of these applications, shedding light on the advancements that have significantly enhanced Threat modeling. Furthermore, the research includes a meticulous gap analysis within the existing literature, revealing areas where further investigation is warranted. Identified gaps in the current research landscape serve as a foundation for proposing future research directions in AI-enhanced Threat modeling. The current literature related to the current landscape of Artificial Intelligence research in cybersecurity predominantly focuses on articles and active studies pertaining to the detection and prevention of cyber-attacks. However, there is a noticeable gap in the existing literature when it comes to leveraging Artificial Intelligence to enhance Threat modeling. While numerous innovative methods have been proposed in recent articles, these predominantly concentrate on Threat modeling within specific domains, such as unmanned vehicles, Cyber-Physical Systems, and Healthcare. There is a lack of generalization in applying these findings to improve Threat modeling practices more broadly. A promising avenue for further research lies in the automation of Threat modeling. Existing literature predominantly emphasizes the study of Threat generation areas, leaving other crucial aspects, such as Architecture representation and Model validation, in need of more comprehensive exploration and analysis.