Generative AI for Agriculture (GAIA) – Phase I & II

GAIA is a two-phase initiative led by IFPRI to enhance agricultural advisory services through Generative AI. Phase I (2023–2024) focused on designing an open-source RAG (Retrieval-Augmented Generation) framework and curating a corpus of context-specific agricultural content from CGIAR and CABI to improve the quality, accuracy, and contextual relevance of AI-generated advisories. Phase II (2024–2027) scales this work by expanding the content corpus to new partners (e.g., FAO, NARES), fine-tuning LLMs for small-scale producers, developing a GenAI ethics and data governance framework, integrating real-time data for dynamic advisories, and building benchmarking protocols. Both phases aim to ensure equitable, trustworthy, and effective AI tools for small-scale producers, especially women and marginalized groups, in Africa and South Asia.