Developing the Cybersecurity Sectoral Risk Index: Computational Analysis of Human-AI Collaboration

By computational analysis based on an adaptive network modeling approach based on self-modeling systems, this paper presents a study of human-AI collaboration in educational and research institutions facing evolving cybersecurity threats. Focusing on stress, urgency, and decision-making under pressure, it introduces a dual-model system combining an employee model with an AI Coach. The model reflects how cognitive, emotional, and contextual factors shape human threat responses and how AI can adaptively support the employee when performance drops. Simulations show that under high workload or stress, employees often struggle to detect and act on cyber threats. That could lead to delayed or absent responses. The AI Coach continuously monitors behavioral patterns and activates support mechanisms when it's necessary. This results in more consistent and accurate responses, even in challenging conditions. These findings suggest that AI-supported adaptive models can strengthen situation awareness and reduce security risks by supporting employees.