AI can help protect forcibly displaced: UNHCR highlights responsible innovation for refugees

GENEVA - From forecasting potential forced displacement movements and assessing damaged buildings after a crisis, to mapping refugee laws and improving supply chains, UNHCR, the UN Refugee Agency, is using artificial intelligence (AI) and advanced analytics to help humanitarian teams prepare earlier, respond faster and better support forcibly displaced and stateless people.

Ahead of the AI for Good Global Summit in Geneva next week, which will bring together humanitarian, technology and development actors to explore how AI can be used responsibly for social good, UNHCR is highlighting practical examples of how AI is already supporting humanitarian action in displacement settings through its Innovation Incubator and Digital & Data Innovation Funds.

"With displacement near record highs and resources shrinking, we have to find smarter ways to protect people," said Hovig Etyemezian, UNHCR's Head of Innovation. "Responsible AI can help humanitarian teams see risks earlier, plan better, respond faster and make limited resources go further."

Several pilot projects in the organization are exploring how predictive analytics, including AI, can support anticipatory planning. In Somalia, for example, one pilot has looked at how AI can help analyse possible internal displacement and refugee movement trends. In Ukraine, agent-based modelling has supported predictive analysis of return movements. Similar approaches are being tested in parts of East Africa, the Sahel, Latin America's mixed movement corridors and climate-affected areas, where recurring displacement and historical data can help inform preparedness.

Beyond anticipatory action, other projects are testing how AI and advanced analytics can support day-to-day humanitarian operations. In Tanzania, predictive tools are helping with pharmacy stock management, while in Burundi, analytics are supporting fuel management. Across supply chains, digital tools are also being piloted to improve the tracing of relief items and strengthen accountability.

Generative AI is also being piloted in support of refugee protection work. One tool enables teams to map and analyse complex domestic laws and policies across more than 200 countries more efficiently to review rights of refugees, internally displaced and stateless people, reducing the time needed for legal research and promoting more consistent analysis.

Other AI-supported tools are being explored to help teams better understand where delays or pressure points may arise in national asylum procedures, including refugee status determination, to support more informed planning in asylum systems under strain. UNHCR underlines that its approach to innovation is grounded in meaningful participation, working with refugee-led organizations and displaced communities to identify challenges and co-design solutions.

"Innovation works best when it is shaped with the people who understand the risks, needs and realities on the ground. Refugees and displaced communities bring knowledge, trust and practical solutions that AI cannot replace. Our aim is to use innovation to support their leadership and strengthen the solutions they are already building."

UNHCR is also clear about the limits and risks of AI. Data gaps, bias, sudden shocks and weak digital infrastructure can reduce the reliability of models, particularly in volatile humanitarian settings. For this reason, UNHCR is adopting AI cautiously and selectively, scaling tools only when they are evidence-based, ethically sound and operationally appropriate. Any models that are deployed will incorporate human verification.

"Innovating with AI also means knowing when not to use AI," said Etyemezian. "There is no one-size-fits-all solution. Sometimes AI can add real value; in other contexts, simpler tools, local knowledge or human judgement will always be more appropriate. Technology must support humanitarian judgment, not replace it. Decisions that affect people's lives must remain in human hands."

UNHCR is also addressing the risks that AI can pose to displaced people, including misinformation, manipulated content and gendered deepfakes. For refugees, access to reliable information can be a matter of safety, especially when false information affects movements, access to services or trust in humanitarian actors.

Marking the 75th anniversary of the Refugee Convention this year, UNHCR stresses that technological progress must serve protection, dignity and solutions. AI can help humanitarian actors be faster, better prepared and more effective, but only if it is grounded in human rights, good governance, transparency, accountability and the meaningful participation of the communities we serve.

To request an interview with UNHCR's experts, including Hovig Etyemezian, UNHCR's Head of Innovation, who will be available during the Summit, please contact:

• Matthew Saltmarsh, [email protected], +41 79 967 9936
• Biel Calderon, [email protected], +41 76 625 4616

For additional information, please contact:

• Eujin Byun, [email protected], +41 79 747 8719

Related news and stories