AI can support humanitarian crisis management. This article explores its potential, ethical challenges, and how MEDAIGENCY is promoting responsible, human-centred solutions for health emergencies across the Mediterranean.
Written by Tamara Mokdad | Humanitarian Engineering Initiative – American University of Beirut
Humanitarian crises are growing in scale and complexity, demanding faster, more effective action. Artificial intelligence (AI) is increasingly recognized as a tool that can support the entire crisis-management cycle – from preparedness and response to recovery – particularly in public health [1, 2]. Evidence suggests that AI can strengthen institutions’ ability to process complex information without replacing institutional responsibility or the human voice in times of crisis [2].
During the preparedness phase, AI has made its clearest contribution to disaster-risk knowledge and early warning. A 2025 systematic review screened 600 records and identified 67 studies relevant to AI-supported early warning systems for infectious diseases. It found that machine learning is increasingly used to combine climate, environmental and surveillance data, helping identify unusual disease trends earlier than manual review [3]. AI, particularly computer vision, is also well suited to recognizing patterns in structured historical data, including flood mapping from satellite imagery and post-disaster damage detection [2]. However, interpreting genuinely novel conditions – a process crisis-management researchers describe as sensemaking – remains a distinctly human capability. In practice, AI can accelerate the recognition of familiar patterns, but it cannot replace expert judgment when a crisis falls outside the historical record.
Predictive modelling is especially valuable for epidemic forecasting. Machine-learning models that draw on historical outbreak data, environmental variables and real-time surveillance can improve forecasting and help authorities anticipate resource needs and target public-health interventions [3, 4]. Similar approaches can support disaster-scenario modelling. The United Nations University Institute for Environment and Human Security describes how AI-driven simulations can help humanitarian organizations and governments stress-test contingency plans and optimize resource allocation before a disaster. It also stresses that these tools should be co-designed with the local actors and communities affected by their outputs [5].
The response phase presents a different set of constraints. Decisions often need to be made quickly and with incomplete information, while the infrastructure on which AI tools depend – including network connectivity – may itself be disrupted [2]. Within these limits, the literature identifies several potential applications: AI-supported early-warning alerts, conversational agents or chatbots for displaced populations, AI-enabled telemedicine, and automated supply-chain logistics [4]. Natural language processing can also translate and synthesize multilingual field observations, news reports and social-media content to help build a real-time operational picture. AI-assisted triage tools have likewise been tested to help clinical teams prioritize care when time and resources are limited [4]. However, much of the available evidence still comes from case reports and pilot deployments. As adoption grows, independently evaluated operational systems will be essential to addressing the remaining evidence gaps.
AI may have its most significant structural value in humanitarian health crisis management by strengthening coordination across borders and between agencies. This contribution can be understood at three levels:
Over the longer term, the tools used to anticipate and respond to crises can also guide investments in resilience, helping recovery efforts not only address damage but reduce the vulnerabilities that existed before the crisis.
The ethical challenges associated with AI in humanitarian and crisis settings cannot be solved by technology alone. Data privacy, algorithmic bias and unequal access to AI tools remain central concerns [4]. Irfan and Sirvent argue that discussions about AI ethics in global health often prioritize professional expertise – including that of clinicians, data scientists and policymakers – over community judgment. They call for greater decision-making power for at-risk communities, including through participatory ethics councils that can shape evaluation metrics, equity requirements and deployment decisions, particularly in low- and middle-income countries [7]. Such participation can strengthen safety, accountability and governance while supporting public-interest AI: systems designed and governed to benefit the public rather than only commercial or institutional interests [7]. This approach is consistent with the view that trust-building, relationship management and contextual judgment remain human capabilities that AI cannot replace [2]. In humanitarian response, AI should therefore complement – not substitute for – human judgment.
Within MEDAIGENCY, the responsible and ethical use of AI has the potential to save lives, reduce the impact of crises on affected populations and strengthen the resilience of health systems across the Mediterranean.
References
[1] Chumachenko, D., Morita, P. P., Ghaffarian, S., & Chumachenko, T. (2024). Editorial: Artificial intelligence solutions for global health and disaster response: Challenges and opportunities. Frontiers in Public Health, 12, 1439914. https://doi.org/10.3389/fpubh.2024.1439914
[2] SAPEA (Science Advice for Policy by European Academies). (2025). Artificial intelligence in emergency and crisis management: Rapid evidence review report. Prepared for the Scientific Advice Mechanism to the European Commission. Munich: SAPEA. https://doi.org/10.5281/zenodo.17737962
[3] Villanueva-Miranda, I., Xiao, G., & Xie, Y. (2025). Artificial intelligence in early warning systems for infectious disease surveillance: A systematic review. Frontiers in Public Health, 13, 1609615. https://doi.org/10.3389/fpubh.2025.1609615
[4] Haykal, D., Goldust, M., Cartier, H., & Treacy, P. (2025). AI in humanitarian healthcare: A game changer for crisis response. Frontiers in Artificial Intelligence, 8, 1627773. https://doi.org/10.3389/frai.2025.1627773
[5] United Nations University Institute for Environment and Human Security. (2024). 5 ways AI can strengthen early warning systems. https://unu.edu/ehs/series/5-ways-ai-can-strengthen-early-warning-systems
[6] World Economic Forum. (2023, February 18). Türkiye-Syria earthquake: How AI and emerging tech are helping relief efforts. https://www.weforum.org/stories/2023/02/turkiye-syria-earthquake-ai-emerging-tech-relief-efforts/
[7] Irfan, B., & Sirvent, R. (2026). Rethinking expertise in artificial intelligence ethics for global health. International Health, 18(1), 134-136. https://doi.org/10.1093/inthealth/ihaf114
Last Update
Next Med