Artificial intelligence (AI) offers potential benefits for agroecology, including early warning systems, crop disease diagnosis, climate information and improved access to agricultural knowledge. However, its contribution to food security and sustainable agriculture depends on how technologies are designed, governed and used. This webinar, organized by the Agroecology Coalition through its Research, Innovation and Education Working Group, brought together representatives from research, civil society, farmer innovation networks, technology initiatives and public authorities to examine whether AI can support agroecological transitions or reinforce existing inequalities. Contributions drew on the High Level Panel of Experts on Food Security and Nutrition’s background note on AI, digitalization and data governance, as well as the IPES-Food report Head in the Clouds, which examines the promises and risks of AI in food systems.

Speakers explored the risks associated with corporate concentration, technological dependency, environmental costs, linguistic bias and the extraction of farmers’ and Indigenous Peoples’ knowledge without adequate consent or benefit-sharing. They emphasized that food-insecure regions often lack the data needed to develop reliable AI applications, increasing the risk of inappropriate recommendations. Examples from the global community of practice Prolinnova, Switzerland’s AgriData.ch initiative and the open-source Agroecology Map illustrate alternative approaches based on locally grounded innovation, secure data exchange and community control. The discussion also suggested that not all knowledge should be digitized and that non-digital solutions may remain more appropriate in some contexts.