About FAO
The Food and Agriculture Organization (FAO) is a specialized agency of the United Nations that leads international efforts to defeat hunger. Our goal is to achieve food security for all and make sure that people have regular access to enough high-quality food to lead active, healthy lives. With over 194 member states, FAO works in over 130 countries worldwide. We believe that everyone can play a part in ending hunger.
Activities by FAO
Animal Feed Balance calculation
Africa EthiopiaBurkina FasoMaliChadThis project aims to use Large Language Models (LLM) and advanced ML techniques to assist farmers in creating the best mix proportions for animal feed based on available ingredients, helping to improve milk production levels and quality.
Activity Type AI Tools/SolutionsTechnical assistanceDetecting Fall armyworm through user submitted photos (FAMEWS)
Africa Not Applicable
The FAMEWS mobile app and online platform monitor Fall armyworm (FAW) infestations. The app uses an image recognition model to provide immediate advice to farmers. It is a free tool helping to reduce crop yield losses and is accessible to a wide audience.
The most important challenge was promoting the adoption of the application and convincing FAO members to share their data. The second challenge was the sustainability of the project as it is difficult to maintain and promote the system without any financial support. The accuracy of the collected data was also another challenge as it is crowd-sourced data.
Toolkit on Digital Transformation for People-Oriented Cities and Communities
Global Not ApplicableA comprehensive guide designed to help cities and communities leverage digital technologies for sustainable development. The toolkit, developed with 13 UN entities, provides practical strategies and tools through 12 modules covering topics from governance to smart manufacturing.
Activity Type Research/Reports/AssessmentsPolicy/Regulatory GuidanceAwareness/Advocacy2025 Geo-AI Challenge
Global Russian FederationUzbekistan
Satellite imagery with increasingly high spatial and temporal resolution is now widely available, and one of its most common applications in agricultural monitoring is the identification of cropland extent. Although a large body of research has focused on cropland mapping using satellite observations, accurate mapping in arid and semi-arid regions remains a major technical challenge because cropland is often difficult to distinguish from pasture and steppe.
The 2025 Challenge selected “Cropland Mapping in Dry Environments” as its central theme and invited data scientists, researchers, and AI practitioners to develop innovative and cost-effective solutions based on artificial intelligence (AI) and machine learning (ML). Two pilot regions, Fergana (Uzbekistan) and Orenburg (Russia), were chosen because they contain representative pasture and steppe landscapes. The challenge was hosted on the Zindi platform, which connects a global community of data scientists.
During the submission period from July 2 to September 25, 2025, a total of 173 valid submissions were received. The accuracies of these submissions ranged from 53% to 91%. The top 10 entries on the leaderboard were selected for further evaluation based on both classification performance and model novelty, and the top three submissions in the second-round evaluation were ultimately awarded prizes.