Sažetak
This paper examines how intelligent automation can support critical infrastructure during crises while preserving human responsibility. Qualitative document analysis compares nine cybersecurity, resilience and artificial intelligence (AI) governance sources against five coverage criteria; four illustrative scenarios inform a domain-specific adaptation of established human–automation research. The comparison distinguishes broad operational guidance from AI-specific oversight and European legal duties. The resulting four-category model separates monitoring, decision support, pre-approved support execution and response requiring operator authorization. NIS 2, the Critical Entities Resilience (CER) Directive and the EU AI Act connect these functions to cybersecurity, all-hazards resilience and AI governance. Meaningful oversight requires evidence review, safeguards against automation bias and maintained operator competence. Agentic AI, medical-device interventions and time-critical protective control are excluded. The contribution is a conceptual synthesis, without empirical performance claims; sector-specific validation remains necessary.
Ključne riječi
Intelligent Automation Critical Infrastructure Protection Crisis Management Cyber- Physical Systems Operational Resilience