The United Arab Emirates has unveiled a groundbreaking artificial intelligence system specifically engineered to address agricultural challenges worldwide. Developed by ai71 in collaboration with 15 global organizations including CGIAR and the Gates Foundation, the AgriLLM platform represents a significant advancement in agricultural technology.
Unlike commercial AI tools such as ChatGPT, AgriLLM operates on an open-access model, allowing free usage, modification, and development by users globally. This initiative addresses a critical gap identified by the UN’s Food and Agriculture Organization, which reports that 75% of family farmers worldwide lack reliable agricultural support systems.
According to Mehdi Ghissassi, Chief Product and Technology Officer at ai71, the system distinguishes itself through specialized training on high-quality agricultural datasets. “While general models like ChatGPT are trained on broad, multi-domain data, AgriLLM is trained on meticulously curated agricultural information from our global partners,” Ghissassi explained.
The system’s effectiveness is demonstrated through rigorous testing, with internal evaluations showing AgriLLM delivers factually correct responses approximately 30% more frequently than GPT-4o when addressing agricultural queries. The model prioritizes accuracy over response length, providing concise, evidence-based guidance rather than potentially misleading comprehensive answers.
AgriLLM’s training incorporates an extensive knowledge base including over 350,000 agricultural documents, 50,000 research papers, and 120,000 validated farming questions and answers. This specialized training enables the system to address crop-specific issues, regional growing conditions, and climate-related challenges that typically challenge general-purpose AI systems.
The platform’s functionality adapts to user specificity, with broad queries generating general advice while follow-up questions regarding soil type, location, or climate conditions trigger increasingly targeted responses drawn from its verified agricultural knowledge base.