Social acceptance of large language models in agricultural extension and advisory: a rapid review

This review examines how large language models are being adopted in agricultural extension and advisory services (EAS). It applies two frameworks, Rogers’ Diffusion of Innovation and Wüstenhagen et al.’s Social Acceptance Framework to analyze 12 studies (via a PRISMA rapid review, coded manually and in NVivo). It finds LLMs are already used for question answering, climate analytics, and pest identification, offering useful agronomic insights but also risking disruption to advisors’ traditional role. Since most LLM development is led by private firms, the review flags concerns about transparency, accountability, and inclusion, and argues that EAS practitioners, through their local knowledge and trusted relationships, are well placed to guide responsible adoption, provided they get targeted capacity-building and inclusive strategies.