AI-Driven Disaster Relief: A Systematic Literature Review of Role-Based Coordination Platforms

The Core Problem Disaster response fails not because help is unavailable. It fails because the right help does not reach the right place at the right time. NGOs, volunteers, government bodies, victims, and donors all work in the same crisis but rarely through the same system. That disconnect has real costs. What This Review Examined Twenty peer-reviewed studies were analyzed across five areas: Crisis informatics and real-time information flow Humanitarian logistics and resource routing Volunteer and NGO field coordination AI-driven resource allocation and prioritization Social media tools for detecting victim needs What the Studies Found The technical work across these domains is solid in parts. But a few problems keep coming up: Logistics models assume clean data and central authority - neither exists during actual disasters AI allocation tools perform well in simulations, poorly in the field Victim distress signals picked up via social media rarely connect to any dispatch mechanism Volunteer deployment stays largely manual, with little cross-NGO visibility Information reaches coordinators and field staff at different times, from different sources The Actual Gap Not one reviewed platform was built to serve more than two stakeholder groups at once. Tools exist for NGO managers, for donor tracking, and for administrative reporting; but nothing brings all five roles into a shared working environment. That fragmentation is where coordination breaks down.