Cyber Security Awareness and Responsible AI Adoption: A Mediation Analysis of AI Risk Perception

Abstract: The rapid integration of artificial intelligence into organisational workflows has introduced unprecedented cybersecurity challenges, necessitating a deeper understanding of how users perceive and respond to AI-related risks. This study investigates the mediating role of AI Risk Perception in the relationship between Cyber Security Awareness and Secure AI Usage Behaviour. Drawing on Protection Motivation Theory and employing a cross-sectional survey design, data were collected from 300 respondents across diverse industries using a structured questionnaire measuring three latent constructs. Mediation analysis, conducted using the PROCESS macro with bootstrap confidence intervals, revealed that Cyber Security Awareness has a strong positive effect on AI Risk Perception (path a: β = 0.712, p < 0.001), which in turn positively influences Secure AI Usage Behaviour (path b: β = 0.444, p < 0.001). The direct effect of Cyber Security Awareness on Secure AI Usage Behaviour remained significant (path c′: β = 0.413, p < 0.001), while the indirect effect through AI Risk Perception was also significant (β = 0.316, 95% CI [0.229, 0.404]). These findings support partial mediation, indicating that AI Risk Perception accounts for approximately 43.3% of the total effect. The results underscore the importance of integrating cybersecurity education with AI risk awareness programmes to foster responsible AI adoption.