How AI can help fund resilience, not disasters

By co-authors from the Global Initiative on Resilience to Natural Hazards through AI Solutions and the Working Group on AI for Climate Applications

In 2024, the World Meteorological Organization (WMO) recorded 617 extreme weather events based on reports from countries worldwide. Of those, 152 were classified as unprecedented and 297 as unusual.

According to WMO’s State of the Global Climate report, these disasters displaced more than 824,000 people, injured 1.1 million, and caused 1,700 fatalities.

Although the numbers are troubling, they do not capture the full spectrum of impacts and consequences as global climate change takes hold. Intensifying disasters also, for example, entail a financial burden.

This hinders development in the most vulnerable – and often most exposed – regions and, in turn, can push climate change mitigation and adaptation measures further out of reach.

Compounding the cost of climate change adaptation are crucially overlooked factors in global and regional finance, such as on the role of ecosystems like wetlands and forests in the provision of essential services.

A recent study on European banks found that ecosystem degradation and biodiversity loss affect financial systems significantly and must be included and integrated into risk assessments.

At the 2025 Pan-European Summit on Climate Resilience, leaders and experts called for urgent, just, and science-based action to strengthen resilience and develop strategies grounded in local communities and natural systems, noting Europe’s position at the forefront of the pathway to a resilient future.

Applying AI to global datasets

This year’s International Day for Disaster Risk Reduction calls on governments and institutions to “Fund Resilience, Not Disasters.” That means financial decisions and assessments must better account for the interplay of climate risks, land degradation and biodiversity loss.

Artificial intelligence (AI) – through its ability to quickly ingest, process, and find patterns within vast, complex datasets – can help flag key factors across and within financial, economic, social, and natural systems.

Machine learning is already being applied to identify hidden drivers of heatwaves and improve the detection and predictability of such extremes. Similarly, AI techniques have increased the predictability of cyclones in the tropics and other regions.

Other studies are using machine learning to identify areas where agriculture requires transformative adaption.

For land degradation, AlphaEarth Foundations is fusing satellite-based Earth observation and climate data to develop detailed global maps of land degradation, deforestation, soil health, and other variables.

Biodiversity experts have used machine learning to model future forest fire susceptibility in monoculture regions and to highlight benefits of mixed forests.

Unpacking complex risks

In terms of financing, AI can help shift the focus from urgent disaster relief to long-term investments in climate resilience. It can do so by tackling the immense complexity of the climate crisis.

Climate threats often consist of multi-hazard events with cascading financial and social effects. AI excels in processing the vast, complex data from environmental, climate, and financial systems. From there, it can create robust, systemic risk models that human analysts could never manage at scale.

Optimizing cash flows

Another way AI can contribute is by optimizing cash flows. Specifically, it can transfer complex climate-biodiversity-land (CBL) data into decision-ready finance metrics, making CBL risk measurable, priceable, and bankable.

For private investors, AI provides strong analytics on ways to prevent losses and, by showing how to stabilize cash flows, can make nature-based solutions more attractive.

By balancing financial returns and environmental benefits, AI lets private investors go from reactive to proactive.

For public resources, AI can determine where funds would work best to de-risk private investments in the context of blended finance strategies. This would create a chance to boost climate and nature resilience and reduce systemic risks.

Making AI accessible and explainable

To fully achieve AI’s potential for sustainable resilience, some critical conditions must be met.

Countries and organizations need shared and widespread data and computational infrastructures, with equitable access to standardized, high-quality data enabling effective results with AI.

Still, people may rightly not accept AI-based decisions with direct social consequences.

To foster trust, AI models need to explain themselves.

Explainable AI (xAI) ensures that models, often trained on decades of expert knowledge, also reveal clearly why they make a certain prediction.

In the case of AI-driven investment and funding decisions, this is key for grounding financial outputs in verifiable physical and expert principles. Thus, AI can help guide capital toward proactive, verifiable resilience projects.

Pulling together

Our call for action on this International Day for Disaster Risk Reduction is clear: governments and businesses must prioritise resilience in funding decisions.

This can be aided by integrating AI into long-term policy and financial frameworks for climate resilience.

Here, international coordination will be crucial for success.

Interdisciplinary and international standardization efforts, like the Global Initiative on Resilience to Natural Hazards through AI Solutions and its new Working Group on AI for Climate Applications, can provide important insights and guidelines.

Together, we must determine how AI can best support informed financial decisions related to the climate and nature crisis.

The Global Initiative on Resilience to Natural Hazards through AI Solutions is co-led by various UN agencies: the International Telecommunication Union (ITU), the UN Environment Programme (UNEP), the UN Framework Convention on Climate Change (UNFCCC), the Universal Postal Union (UPU), the World Meteorological Organization (WMO), and the UN Educational, Scientific and Cultural Organization (UNESCO).

The Working Group on AI for Climate Applications is a collaborative expert group within the Global Initiative that aims to harness AI for practical climate change mitigation and adaptation, particularly addressing the needs of developing countries, least developed countries, and small island developing states.

This article was prepared by the following co-authors in collaboration with the two groups:

• Monique Kuglitsch (Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, Berlin, Germany)
• Kamal Kishore (United Nations Office for Disaster Risk Reduction, Geneva, Switzerland)
• Arthur Hrast Essenfelder (European Commission Joint Research Centre, Ispra, Italy)
• Zikun Lyu (Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, Berlin, Germany)
• Jorge Pérez-Aracil (University of Alcalá, Madrid, Spain)
• Jennifer Selby (Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, Berlin, Germany)
• Andrea Toreti (European Commission Joint Research Centre, Ispra, Italy)
• Elena Xoplaki (CMCC Foundation Euro-Mediterranean Center on Climate Change, Bologna, Italy)

The views expressed in this article belong to the co-authors and may not, in any circumstances, be regarded as stating an official position of institutions with which they are affiliated.

Header image credit: Adobe Stock