About IFAD
At the International Fund for Agricultural Development (IFAD) we invest in rural people, empowering them to increase their food security, improve the nutrition of their families and increase their incomes. We help them build resilience, expand their businesses and take charge of their own development.Activities by IFAD
Global Initiative on AI for Food Systems
Global GermanyIndiaThe Global Initiative on AI for Food Systems, is a collaborative global effort led by the International Telecommunication Union (ITU) together with the Food and Agriculture Organization (FAO), World Food Programme (WFP), International Fund for Agricultural Development (IFAD), and other partners to harness the transformative potential of artificial intelligence (AI) and digital technologies across the entire food value chain. It aims to boost agricultural productivity, strengthen food security, and build resilient, sustainable food systems by embedding AI into areas such as precision agriculture, supply chain optimization, and real-time decision support while addressing climate and resource challenges. A central objective is to establish common frameworks, shared digital infrastructure and standards that ensure interoperability, security, and adaptability of AI applications, enabling governments, innovators and smallholder farmers, who produce a significant portion of global food but often lack connectivity and support—to benefit from AI innovation at scale. The initiative supports knowledge exchange, synergistic partnerships, and proof-of-concept deployments to demonstrate scalable, responsible, and inclusive AI solutions that contribute to global food system resilience.
Activity Type AI Tools/SolutionsTrainings/WorkshopsTechnical AssistanceResearch/Reports/AssessmentsPolicy/Regulatory GuidanceAwareness/AdvocacyNetworks/Mentorship/ExchangeFlash Flood Early Warning System (FFEWS)
Asia and the Pacific Bangladesh
The Flash Flood Early Warning System (FFEWS) was developed under the Haor Infrastructure and Livelihood Improvement Project / Climate Adaptation and Livelihood Protection (HILIP/CALIP) project and further scaled under Promoting Resilience of Vulnerable through Access to Infrastructure, Improved Skills, and Information Project (PROVATi3) projects. These projects in Bangladesh funded by IFAD represent an important application of Artificial Intelligence (AI) and particularly Machine Learning (ML) models for climate resilience and disaster risk management. The system combines machine-learning-enhanced hydrological and hydrodynamic models with real-time environmental monitoring to improve the prediction of flash floods and flood inundation in some of Bangladesh’s most climate-vulnerable regions.
The solution also incorporates IoT-style monitoring infrastructure and multi-channel dissemination systems, including SMS alerts, voice messaging, mobile applications, digital display boards, and social media platforms. Through this integrated approach, the system generates near real-time flood forecasts and location-specific warnings that support community preparedness, disaster response, and protection of livelihoods. The initiative demonstrates how AI-enabled predictive analytics and geospatial technologies can strengthen climate adaptation and anticipatory action for vulnerable rural communities.
Garden AI Benchmark
Global Not Applicable
AI models, including Large Language Models (LLMs), are essential building blocks for many digital solutions and services. As these technologies advance, choosing the right model is key to ensuring reliable and effective results. The Garden AI Benchmark is a collection of IFAD-specific tests for AI models, with automation scripts for testing models at scale and an interactive dashboard to evaluate results in a clear and easy-to-understand manner. The benchmark allows for the assessment of different models across several important areas for the IFAD context, using standardized evaluation metrics to measure both accuracy and ability to handle different tasks. The benchmark helps IFAD make better-informed model selection choices for specific purposes across a variety of solutions and use cases.
Current model tests include:
● Choice Selection
● Language Translation
● Information Mapping
● Information Retrieval
● Complex Reasoning
● Hallucination Abstention
● Audio Transcription
Current model types include:
● Embeddings
● Generation
● Reasoning
● Translation
● Speech-to-Text
● Text-to-Speech
Workable Intelligence Service Hub (WISH)
Global Not ApplicableWISH is a central orchestration engine to define, compose, and execute AI-driven workflows via reusable prompts, bundles, and behaviors. WISH is a foundational component of IFAD's Enterprise AI Architecture, allowing for the seamless integration of AI features and capabilities in IFAD's corporate systems for project procurement, financial management, portfolio operations, and more.
Activity Type AI Tools/SolutionsInfrastructure/Systems Development