AI in Disaster Resilience: Bridging Science, Standards and Innovation

Good morning, Excellencies, colleagues, and  distinguished guests.

Let me begin by thanking ITU and partner organizers for bringing together leaders, experts, innovators, policy makers and representatives from the AI, disaster resilience and scientific communities into one room today. We need exactly this kind of dialogue if we are to ensure that AI delivers tangible benefits for people and communities around the world.

As many of you know, WMO's Members are the National Meteorological and Hydrological Services that provide the authoritative forecasts and early warnings on which governments rely to protect lives and livelihoods. We therefore recognize the enormous potential of AI to strengthen this mission.

AI has changed what is technically possible. AI is transforming how we produce, communicate and use forecasts, with enormous potential to improve warnings of tropical cyclones, flash floods and sand and dust storms, while strengthening climate services for sectors such as agriculture, health and energy.  

But disaster management is not just about prediction.  It is about enabling better decisions that protect and save lives.

Allow me to make three key points.  

First, AI offers a major opportunity to accelerate progress under Early Warnings for All by making climate intelligence more timely, more accurate and more accessible.  

One of the most exciting developments is that many emerging AI forecasting systems can now produce high-quality forecasts with significantly lower computational requirements than traditional numerical weather prediction.  

This has the potential to democratize forecasting by enabling many more countries to generate and use their own forecasts as the basis for life-saving early warnings.

Second, AI is only as strong as the observations, data and scientific infrastructure on which it is built. Investment in AI must therefore go hand in hand with sustained investment in observing systems, data exchange, computing capacity and human expertise. This is particularly important in Least Developed Countries and Small Island Developing States, where these gaps remain most significant.

Third, AI should support and strengthen National Meteorological and Hydrological Services and not replace them. People trust official warnings because they come from authoritative public institutions that understand local risks, communicate uncertainty responsibly and remain accountable to the communities they serve.  

AI should enhance that public service mission, not bypass it.

These principles are already guiding WMO's work on AI.

Colleagues,  

This summit marks the launch of our report “Leveraging AI to enhance multi-hazard early warning systems”, which demonstrates that AI is opening a new chapter in hazard detection, monitoring and forecasting.  

A concrete example is the WMO-led Artificial Intelligence for Nowcasting Pilot Project to develop, test and evaluate AI-based methods for severe weather nowcasting in targeted developing countries.  

There are many other promising examples in flood forecasting, water resources management, and sub-seasonal prediction. 

In preparation for the UN Global Dialogue on AI Governance, WMO convened a consultation with its Members  and one message emerged consistently across our Members: equity of access must remain central if AI is to strengthen nationally owned, authoritative and trusted early warning services.

To support this work, WMO has also established a Joint Advisory Group on Artificial Intelligence to guide the responsible integration of AI into operational weather, climate and water services. Let us work together to ensure that AI serves as a powerful tool for early warnings, not only by advancing technology, but by strengthening trust, scientific integrity, and international cooperation.  

Together, we can ensure that no country and no community is left behind in benefiting from AI-powered early warning systems that protect lives and livelihoods.

Thank you.