From Data Points to Designers: Climate Agronomy Learning Commons for Kenya's Western Highlands

This research proposes a Climate Agronomy Learning Commons to address climate vulnerability among smallholder farmers in Kenya’s Western Highlands (Vihiga, Kakamega, Bungoma). It critiques current agricultural extension systems as generic, underfunded, and exclusionary, while digital farming platforms often exploit farmer data without transparency ("ghost tech loops"). The solution integrates AI, para-agronomist networks, and a Community Agronomy Data Trust (CADT) to democratize climate-smart agriculture. The model emphasizes local accountability, multilingual voice/USSD accessibility, and farmer co-design to combat digital exclusion, governance gaps, and climate shocks. Key innovations include: AI as a co-designed amplifier (not an oracle) for hyper-local agronomy advice. Legally recognized CADT ensuring data sovereignty, consent, and benefit-sharing. Triple-arm pilot (AI-only, hybrid, human-only) to assess trust and inclusivity. Policy integration via a County-Level AI Readiness Index and farmer-driven governance.