AI for early warnings for all: From innovation to impact

This session marks the official launch of the AI for EW4All Group report and explores how artificial intelligence can strengthen people-centred multi-hazard early warning systems across the full MHEWS value cycle. Through keynote insights, case studies, and interactive discussion, participants will examine practical applications of AI, lessons learned, and opportunities to scale implementation at country level.

Workshop opening remarks and official launch of the 'Leveraging AI to Enhance Multi-Hazard Early Warning Systems (MHEWS): A practical resource to support Early Warnings for All (EW4All)' report co-authored by ITU, WMO, UNDRR and IFRC. Followed by official report launch photograph.

Presentations highlighting key findings and recommendations from the AI for EW4All Group report across the four pillars of the MHEWS value cycle and interpillar synergies.

Case study presentations featured in the AI for EW4All Group report showcasing operational AI applications and implementation experiences across the MHEWS value chain.

Interactive audience discussion using Mentimeter focused on scaling AI implementation at country level; capacity, governance, and operational challenges; and priority AI applications across the MHEWS value cycle.

This session will discuss interpillar insights, the latest progress from the AI for EW4All Group, highlighting pilot projects and country-level implementations of AI-driven tools that strengthen people-centred MHEWS. Speakers will share practical experiences, technical lessons, and key insights, while exploring challenges, gaps, funding and opportunities to scale AI innovations for improved disaster preparedness, response, and global impact.

Presentations will showcase country pilots and applied work featured in the AI report and AI Solutions Catalogue, highlighting practical applications of AI across the MHEWS value cycle. The session will also include a brief overview of an upcoming ITU-GEO Secretariat pilot project in Kenya on extreme heat and digital infrastructure resilience.

An overview of the Multi-hazard, Alert, Zero-gap, Universal (MAZU) framework and its integrated approach to governance, infrastructure, and AI across the MHEWS value cycle. The presentation will explore how China is operationalising coordination across institutions and technologies to strengthen end-to-end early warning capabilities at scale.

This session will highlight the successful completion of the AI for EW4All pilot project in Malawi, supported by WMO CREWS, showcasing how MET Norway, ECMWF, and Malawi’s Department of Climate Change and Meteorological Service (DCCMS) collaborated to develop an AI-based weather prediction system tailored to national needs. The discussion will draw lessons on capacity building, institutional partnerships, and pathways for scaling AI solutions across the weather forecasting and early warning value chain, particularly in Least Developed Countries.

Flash floods are among the world’s deadliest natural hazards, but predicting them globally has long been hindered by a lack of reliable field data. To bridge this gap, Google Research launched Groundsource—an innovative methodology that leverages the Gemini Large Language Model to transform unstructured global news reports into structured, historical disaster data. The project has generated a large-scale, open-access dataset tracking 2.6 million urban flash flood events across multiple countries. By leveraging this unprecedented information, Google can now deliver AI-driven flash flood predictions for urban areas up to 24 hours in advance. In partnership with governments, international organizations, and NGOs, this breakthrough aims to strengthen global climate resilience and help millions stay safe during extreme weather events.

An overview of financing pathways and partnership models that can help accelerate the integration and scaling of AI-enabled solutions across the MHEWS value cycle, with a focus on practical opportunities for country-level implementation and long-term sustainability.

Panel discussion exploring how the four pillars interact across the MHEWS value cycle, highlighting opportunities for integration, collaboration, and scaling, while also examining persistent gaps, operational challenges, and areas requiring stronger coordination to advance AI-enabled MHEWS.

Key takeaways from the day’s discussions, emerging priorities for AI-enabled MHEWS, and the importance of continued collaboration to scale practical, people-centred AI solutions across the MHEWS value cycle.