AI Cybersecurity Capabilities and IoT Resilience: A Moderated Serial Mediation Model of Threat Detection Effectiveness and Incident Response Agility

Abstract

While strategically deploying Artificial Intelligence (AI) is critical for the security of Internet of Things (IoT) ecosystems, organizations often face challenges in converting their AI investments into operational resilience. This study investigates a moderated serial mediation model, shedding light on how AI Cybersecurity Capabilities contribute to achieving IoT Resilience. Based on Resilience Engineering and Dynamic Capabilities theory, we hypothesize that AI Cybersecurity Capabilities as a reflective-formative higher-order construct consisting of Sensing, Learning, Automation, and Decision-Support can increase IoT Resilience by first supporting the operation of the Threat Detection Effectiveness (TDE) and then the Incident Response Agility (IRA) capabilities. Moreover, we hypothesize that Cybersecurity Governance (CSG) is a meta-capability with a unique moderating effect on each phase of the process. The results support the serial mediation pathway (β = 0.10, p < 0.001, 95% CI [0.06, 0.15]) based on the survey data obtained from 327 senior cybersecurity executives in critical infrastructure sectors and analyzed using the Partial Least Squares Structural Equation Modeling (PLS-SEM) with repeated indicators approach. Importantly, CSG moderates the indirect effect, and the serial mediation has a significant stronger effect in high-governance contexts (Index of Moderated Serial Mediation = 0.08, CI [0.04, 0.13]). This research provides a granular process-based cybersecurity value realization model, with suggestions and insights for Security Operations Centers and security governance committees at the board level.