Artificial Intelligence-Based Early Warning System for Disaster Management: A Literature Review Systematic and Bibliometric Analysis

The increasing frequency and intensity of natural disasters globally demands the development of more accurate and responsive Early Warning Systems (EWS). In recent years, Artificial Intelligence (AI) has been increasingly applied in natural disaster mitigation, but the approaches used are still diverse and spread across various domains. This study aims to present a systematic literature review on the application of AI and deep learning in natural disaster early warning systems. This review was conducted following the PRISMA 2020 guidelines by analyzing literature published during the 2020–2025 period. The selection process resulted in 102 studies meeting the inclusion criteria, with 30 full-text articles being analyzed in depth to map disaster types, AI methods, data sources, and characteristics of early warning systems developed in various regions, including Asia and Africa. The review results show the dominance of deep learning approaches, particularly time series-based models such as Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU), particularly in flood forecasting and land deformation prediction. More advanced architectures, such as Transformer, are beginning to be adopted to capture long-term temporal patterns, while the combination of convolutional neural networks (CNN) with remote sensing data is widely used for spatial mapping of disaster events. Furthermore, the integration of sensor data and the Internet of Things (IoT) shows potential in supporting more responsive early warning systems. However, most research remains limited to the modeling or simulation stage, with little discussion of the real-time and operational implementation of EWS. This review highlights the gap between AI model development and the implementation of reliable early warning systems and provides a conceptual foundation for the future development of more integrated AI-based disaster mitigation systems.