This is the authors' abstract. We don't add key points for this paper.
This study investigates the applicability, practicality, and effectiveness of a low-cost AI foundational model (FM) in agricultural extension through the development, fine-tuning, and evaluation of a custom GPT named Utah PeachBot, built using OpenAI’s GPT platform. The research focused on facilitating real-time, evidence-based advisory service support for Extension agents assisting small-scale peach producers in Utah. Methods involved training the GPT with curated, research-based horticultural resources and assessing model outputs through an expert panel of six Extension agents. Results showed high reliability and accuracy for general inquiries about peach cultivation. However, inconsistencies in regional specificity and the practicality of recommendations emerged as limitations. Feedback indicated a need for iterative fine-tuning of the model through continuous expert feedback and integration of local, context-specific data. Recommendations include a phased approach to implementing customized GPTs in agricultural advisory services to improve information dissemination, decision-making quality, and operational efficiency within extension systems.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Hill et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: