Artificial intelligence (AI) has the potential to improve disaster early warning systems around the world, but experts say its success will depend on reliable data, strong governance, and continued investment. That is the key finding of a new report, Leveraging AI to Enhance Multi-Hazard Early Warning Systems, written and coordinated by the International Telecommunication Union (ITU) with chapter contributions from the United Nations Office for Disaster Risk Reduction (UNDRR), the World Meteorological Organization (WMO), and the International Federation of Red Cross and Red Crescent Societies (IFRC). The report was developed under the guidance of the AI for Early Warnings for All (EW4All) Group.

AI Can Strengthen Every Stage of Early Warning Systems

The report examines how AI is being used across the four pillars of Multi-Hazard Early Warning Systems (MHEWS). These pillars include disaster risk knowledge, hazard monitoring and forecasting, warning dissemination and communication, and preparedness and response.

According to the assessment, AI can process large volumes of data faster than conventional methods. This can improve hazard detection, forecasting, and decision support. However, the report stresses that AI is an enabling technology rather than a replacement for human expertise. Effective disaster management still depends on trained professionals, strong institutions, and clear governance frameworks.

The report also identifies operational gaps where AI can improve performance while supporting existing early warning systems.

Infrastructure, Governance, and Equity Are Essential

The report states that AI systems require robust observational infrastructure to deliver reliable results. Ground-based monitoring networks, satellites, weather stations, and in-situ sensors provide the data needed for accurate forecasts. Expanding this infrastructure is considered especially important for Small Island Developing States (SIDS), Least Developed Countries (LDCs), and Landlocked Developing Countries (LLDCs), where monitoring capacity remains limited.

The authors recommend establishing national AI focal points, maintaining human oversight for life-safety decisions, and creating clear accountability frameworks. The report also calls for AI systems that are transparent, multilingual, accessible in low-connectivity environments, and designed with direct participation from affected communities to ensure equitable access.

Long-Term Investment Needed to Scale AI Solutions

The report recommends building early warning systems using interoperable and modular digital architectures. This approach would allow information from risk assessments, forecasting, communication, and preparedness activities to reinforce one another through continuous feedback.

It also notes that many AI applications remain in pilot phases. Moving them into full operational use will require sustained public and private investment, institutional capacity building, and partnerships among governments, research institutions, humanitarian organizations, and the private sector.

The report concludes that future efforts should focus on closing existing gaps in early warning systems while integrating AI responsibly, transparently, and with human oversight. It says that technology alone cannot reduce disaster risk and must be supported by strong institutions, quality data, and inclusive policies to deliver effective protection for communities.