That is the question the Directorate-General for European Civil Protection and Humanitarian Aid Operations (DG ECHO) asked the Scientific Advice Mechanism (SAM) – an EU initiative that provides independent scientific evidence and policy recommendations and brings together leading scientists and European academies.
Their answer was published last December in the AI in Emergency and Crisis Management Rapid Evidence Reviewopens in new tab. Rather than presenting a list of AI tools applicable to disaster and crisis management, the report sets out guiding principles for the selection and use of AI tools. This approach empowers users by providing evaluation criteria to assess whether a particular AI system is fit for purpose. It also highlights case studies on the use of AI for tasks related to disinformation detection, weather forecasting or disaster response. The report concludes with a set of policy options designed to stimulate a discussion on how to strengthen crisis management AI capabilities in Europe.
So, where does AI actually help? The evidence shows that AI performs best in tasks that are well-defined and data-intensive. AI systems can process large volumes of data from many different sources such as satellite imagery, sensor networks and social media, at a pace and scale unseen before. In emergencies, this can help improve early warnings or support rapid damage assessment. For instance, in weather prediction, AI has already outperformed traditional numerical models in some contexts.
However, AI systems still struggle when faced with new and rare situations, particularly where training data is limited. AI’s performance also degrades when systems trained in one context – for example, a specific region – are applied elsewhere. This is especially relevant where hazards, such as wildfires, may occur further north than they did previously. Ultimately, AI is only as good as the data it learns from and operates on. The authors argue that better data preparedness and cooperation are essential for AI to deliver real value.
AI also lacks the ability to understand complex contexts in the way humans do. For this reason, the report emphasises that AI cannot replace human decision-making. The report addresses legal and ethical concerns when using AI in emergencies and stresses the importance of complying with the AI Act and the General Data Protection Regulation (GDPR).
Overall, the report provides a clearer understanding of where AI adds value and what to consider when integrating it into civil protection work.
The Knowledge Network editorial team is here to share the news and stories of the Knowledge Network community. We'd love to hear your news, events and personal stories about your life in civil protection and disaster risk management. If you've got a story to share, please contact us.