Geneva funds AI for humanitarian research checks and local grant reporting

On 21 October, the Swiss Data Science Center and EPFL AI Center announced two funded projects under the Geneva AI Initiative.[1]

Checking the evidence behind aid decisions

At IMPACT Initiatives, an assistant built with a large language model will help reviewers check the methodological soundness and consistency of humanitarian research outputs. The aim is to accelerate quality review for time-sensitive decisions.[1]

IMPACT says it publishes more than 2,000 information products a year. Its REACH initiative supplies data and analysis for humanitarian decisions before, during and after emergencies; PANDA covers programme design, assessments and evaluation. This is work in which a research finding has to travel from a crisis-affected community to organisations deciding how to respond.[2]

The organisation’s 2025–2030 strategy calls for research quality to be embedded across evidence generation, analysis and dissemination.[3]

From paper records to reporting data

GCERF’s CLARO project addresses a different task. It will use locally deployed AI to convert photographs of attendance lists and receipts into structured reporting data. SDSC describes the approach as privacy-preserving.[1]

GCERF—the Global Community Engagement and Resilience Fund—finances community-led civil society programmes that address the local social, economic and political drivers of violent extremism. Its work includes rebuilding trust between communities and the state and supporting alternatives to violent groups. Reporting is part of how a fund follows what those local programmes are doing.[4]

SDSC scientists and machine-learning engineers will support both teams. The announcement funds development; it provides no completed deployment or measured time savings.[1]

Report an error or suggest a correction