From Simulation to Healthcare: KINAITICS' AI Framework for Cyber-Physical Security.

The increasing integration of Artificial Intelligence (AI) into Cyber-Physical Systems (CPS) presents complex cybersecurity challenges, necessitating a reevaluation of traditional threat assessment. The KINAITICS project addresses these evolving threats by conducting in-depth research into cyber-kinetic attacks, where malicious cyber activities manifest as real-world physical disruptions. The project is also dedicated to developing resilient, AI-driven defense mechanisms. This paper outlines KINAITICS' foundational work, including the creation of a tailored KINAITICS Threat Matrix (KTM). This innovative framework systematically identifies, categorizes, and assesses threats unique to AI-integrated CPS. The paper details the KTM's practical application across five high-stakes use cases, ranging from safeguarding nuclear facility simulations to protecting electronic health record (EHR) systems from sophisticated phishing attacks. A central focus of the KINAITICS project is the rigorous development and evaluation of both offensive and defensive AI tools. These tools are designed to investigate, understand, and mitigate the multifaceted threats posed by cyber-kinetic adversaries. The overarching objective is to significantly enhance the resilience of critical infrastructures against advanced cyber-physical threats, ensuring the continued safety, security, and operational integrity of systems vital to modern society.