AI Forecasting for Refugee Crises

An AI-driven tool forecasts displacement patterns in Somalia by analyzing data on conflict, climate, and the economy.

💖 This week's byte: An AI-driven tool forecasts displacement patterns in Somalia by analyzing conflict, climate, and economic data. It allows humanitarian organizations to anticipate resource needs and respond proactively to refugee crises.

📊 Did You Know?

More than one in every 69 people on Earth were forcibly displaced at the end of 2023, the UNHCR reported. The number has increased for 12 consecutive years.

📖 The Story

Who, When, and Where — Context

In mid-2017, UNHCR's (the United Nations Refugee Agency) Innovation Service started focusing on the dataset of forced displacement in Somalia.

Why — Challenge

There are multifaceted reasons behind forced displacement. For example, food insecurity, violent conflict, and gaps in access to food due to price, distance, and instability may cause people to move. The complexity makes forced displacement unpredictable, leaving humanitarian organizations and host communities unprepared to support those impacted.

What and How — Tech Solution

UNHCR designed Project Jetson, a data-driven predictive model that forecasts the displacement of persons in Somalia. This experiment aims to identify, understand, and quantify the specific factors that may be contributing to the forced displacement of Somalis. The experiment demonstrates the possibility that humanitarian organizations can be more proactive in their response efforts, such as improving relief services.

💡 Key Insights

  • 🌍️ Forced displacement is a complex and systemic issue with no obvious cause or solution.

  • 🤖 Mathematical models and machine intelligence can be a game changer in humanitarian sectors, making the conventional invisible visible.

  • 📊 Having access to accurate and quality data is essential to digitalize humanitarian efforts.

✅ Try This

See how to predict forced displacement at:

💭 Share your thoughts: What are the limitations of algorithmic forecasting? How can we (humans) intervene in the process to minimize the potential risks of using AI algorithms?

Reply

or to participate.