What's Happening?
A global scoping review of artificial intelligence (AI) applications in humanitarian contexts reveals that AI is increasingly being explored for disaster monitoring, prediction, situational awareness, crisis informatics, search and rescue (SAR), and coordination.
The review, which synthesized 74 records from scholarly publications and grey literature, indicates a rapid growth in AI-related evidence, particularly in sudden-onset disasters and preparedness or response activities. Most AI applications identified were prototypes or retrospectively validated tools, with limited field implementation or institutional integration. While AI models like deep learning, classical machine learning, hybrid optimization-based systems, and natural language processing are being applied, their operational deployment and real-world evaluation remain insufficient. The study highlights that governance, participation, and equity considerations are inconsistently addressed, with scholarly works focusing on technical aspects and grey literature discussing broader ethical and practical concerns.
Why It's Important?
The findings underscore a critical gap between the technical development of AI solutions and their practical, operational implementation in humanitarian settings. This 'implementation gap' means that potentially life-saving technologies are not reaching those who need them most effectively. The concentration of AI applications in sudden-onset disasters, while beneficial, leaves protracted crises, displacement, public health emergencies, and humanitarian logistics comparatively underexplored. This oversight is significant as many humanitarian crises are long-lasting and require sustained support beyond immediate response. Furthermore, the inconsistent addressing of governance, accountability, and community participation raises concerns about the ethical deployment of AI, particularly regarding data privacy, protection of vulnerable groups, and potential biases. Without robust frameworks for these aspects, the benefits of AI could be undermined, and its application might inadvertently exacerbate existing inequalities or create new challenges for affected populations.
What's Next?
Future research and development in humanitarian AI should shift focus from purely technical development to implementation science, operational evaluation, and governance research. There is a clear need for more prospective field studies, user-centered validation, and participatory design approaches to ensure AI tools are effective and relevant in real-world humanitarian settings. Researchers should prioritize applications that support recovery, resilience, protracted crises, displacement settings, and public health emergencies, which are currently underserved. Policymakers and humanitarian organizations must embed governance, accountability, and community participation from the outset of AI initiatives. Multidisciplinary collaboration is essential to align AI innovation with humanitarian values and operational realities, ensuring that AI meaningfully supports decision-making and improves outcomes for affected populations.
Beyond the Headlines
The review points to deeper implications regarding the ethical and societal dimensions of AI in humanitarian aid. The limited involvement of affected communities in the design and validation of AI systems, despite the use of local data, highlights a potential disconnect between technological solutions and the lived experiences of those they aim to help. This raises questions about true 'human-centered' design and the risk of imposing solutions that do not fully account for local contexts, languages, and cultural nuances. The lack of explicit discussion on formal governance frameworks for responsible AI deployment, particularly concerning international humanitarian law and accountability mechanisms for AI-assisted decisions causing harm, indicates a nascent stage in ethical integration. Addressing these issues will require a fundamental shift towards more inclusive, transparent, and accountable AI development processes that prioritize the rights and dignity of vulnerable populations, moving beyond mere technical feasibility to ensure equitable and just outcomes.
