The 3rd meeting of the Global Initiative on Resilience to Natural Hazards through AI Solutions, which followed on from the workshop the previous day, focused on turning strategic discussions into practical next steps.
In his opening remarks, Deputy Head of Unit Gaetano Vivo (DG ECHO) emphasised the importance of strengthening the links between research, policy and operational practice to advance the responsible use of AI in disaster risk management.
Vivo presented the European Commission’s contribution, developed by different services, outlining three priority areas for integrating advanced technologies into disaster management: reinforcing operational capabilities and technological sovereignty, supporting decision-making in a changing climate, and promoting AI for the public good through international partnerships and capacity-building.
He also highlighted key challenges that must be addressed to ensure trustworthy and effective AI uptake. These include technological barriers such as data availability and system validation, ethical requirements under the AI Act, and institutional needs such as stronger AI literacy and end-user engagement. While initiatives like Destination Earth and the Copernicus Emergency Management Service show significant progress, Vivo underlined that further investment in transparency, usability and de-risking innovation remains essential to deliver real benefits for communities exposed to natural hazards.
The contribution sessions highlighted progress across several of the Initiative’s working groups, including Data, Modelling, and Educational Materials. Updates covered sustainability, AI-for-Good approaches, standardisation, the development of an alerting lab, and new educational resources tailored to least developed countries. Joint activities, such as the Digital Twin Webinar Series, demonstrated growing alignment between data, modelling, and user needs, while comparisons of best practices and questionnaire analyses provided valuable insights into operational challenges and emerging opportunities.
Several concrete proposals underscored the increasing maturity of AI applications for disaster resilience. A notable example came from the European Space Agency, presenting a data-driven global storm-surge and coastal-inundation forecasting approach that attracted strong interest from participants and highlighted the need for robust datasets and improved cross-border data sharing.
Across the discussions, participants reaffirmed that end-user involvement and practical applicability remain essential to ensuring that innovation translates into tangible improvements for disaster risk management.
