Cambridge Fine-Tunes AI Language Model to Automate Wheat Disease Early-Warning Advisories Across 8 Countries

University of Cambridge Department of Plant Sciences researchers (Jacob Smith, Lawrence Bower and Chris Gilligan) announced on August 27, 2026 that they are fine-tuning a large language model on five years of expert-edited advisories to automatically generate concise, localized wheat disease risk summaries for the Disease Early Warning Advisory System (DEWAS) — a CIMMYT-led collaboration of 23 research and academic organizations operating across Bangladesh, Ethiopia, Kenya, Nepal, Pakistan, Tanzania, Zambia and Bhutan. The AI-generated advisories integrate real-time disease forecasts and field survey data to speed dissemination of wheat rust and blast warnings to national plant protection agencies, aiming to give the tens of millions of smallholder farmers across East Africa and South Asia who depend on wheat DEWAS earlier notice to act before outbreaks spread.