Artificial Intelligence for safety and resilience in airport Transportation Systems: A systematic review of Operational, Security, and environmental risks

Airports are critical nodes in transportation systems and operate as safety–critical, tightly coupled socio-technical environments in which operational, environmental, security, occupational, and human-factor risks interact and can cascade into disruptions that test both safety performance and resilience. This study conducts a PRISMA-guided systematic review, supported by bibliometric analysis, of 81 Scopus-indexed studies published between 2019 and 2025 to synthesize evidence on airport risk management and the role of artificial intelligence (AI) across the risk-management cycle, including hazard identification, risk assessment, and mitigation. The review bridges transportation research perspectives spanning policy, systems operation, and safety behavior. The literature is fragmented and tool-centric. Simulation-based studies and single-airport case analyses dominate, while limited translation is observed into governance mechanisms aligned with ICAO Safety Management Systems (SMS). Methodological maturity is stronger for operational and environmental risks, whereas security, occupational, and human-factor risks remain less developed and weakly connected to predictive, system-level safety frameworks. AI shows promise for real-time sensing and forecasting, including foreign object debris detection, pavement-condition monitoring, runway safety prediction, weather-impact forecasting, and surface-movement risk assessment. However, studies frequently rely on narrow datasets, rarely quantify uncertainty and robustness, and seldom address deployment within operational workflows, limiting practical value for transportation safety management. A research agenda is proposed that embeds AI within SMS through explicit operational triggers, enables multi-airport validation via privacy-preserving data sharing, and establishes unified taxonomies, interoperable data standards, and explainable, auditable model requirements to support scalable, governance-ready decision support.