Artificial Intelligence Methods in the Monitoring and Forecasting of Emergency Situations: a Systematic Review and Research Gaps (2019–2026)

This article presents a systematic analysis, based on Scopus-indexed sources, of the application of artificial intelligence (AI) methods to the monitoring and forecasting of emergency situations (ES) over the period 2019–2026. Machine learning and deep learning methods are synthesized across natural hazards (earthquakes, floods and flash floods, landslides) and technological hazards, as well as across the directions of data fusion and digital twins. The analysis identifies four principal research gaps: the lack of multi-hazard approaches, the weakness of rigorous benchmarking against mature methods, the limited capacity for quantitative forecasting of consequences, and the absence of integration into a national platform. The results define the scientific rationale for the dissertation topic and the gap it is intended to fill.