World Bank (2025) Report (multilateral / development partner)

AI-Powered Refugee Forecasting: Preparing for Refugee Movements Before They Happen

World Bank

Ref: SRC-003-UGA-001

https://www.worldbank.org/en/topic/fragilityconflictviolence/brief/ai-powered-refugee-forecasting-preparing-for-refugee-movements-before-they-happen

Accessed: 3/24/2026

Summary

Describes the AI model analysing 90+ variables to forecast refugee arrivals 4-6 months ahead. Key drivers: conflict, economic activity, climate, food prices, and language patterns. Tested using UNHCR data 2014-2023 for South Sudan and DRC arrivals. Achieved high predictive accuracy. Notes model is adaptable for other development challenges.

View Harvard reference

World Bank (2025). 'AI-Powered Refugee Forecasting: Preparing for Refugee Movements Before They Happen'. Washington, DC: World Bank. Available at: https://www.worldbank.org/en/topic/fragilityconflictviolence/brief/ai-powered-refugee-forecasting-preparing-for-refugee-movements-before-they-happen (Accessed 24 March 2026).

Attached File

FILE

SRC-003-UGA-001_AI-Powered_Refugee_Forecasting_Preparing_for_Refugee_Movements_Before_They_Happe.html 44 KB

Download

Linked Cases (1)

UGA-001 UGA Operational Deployment (Limited Rollout)

AI-Enabled Refugee Forecasting for Uganda's Displacement Crisis Response Mechanism (DCRM)

Social assistance Classical ML