Developing and integrating trust modeling into multi-objective reinforcement learning for intelligent agricultural management

• AI-driven precision agriculture boosts efficiency and sustainability. • Reinforcement learning surpasses traditional farming methods. • Human-AI interaction improves trust in AI farm management. • Novel trust model integrates farmers’ real-world experiences. • Embedding trust into RL ensures practical AI recommendations. Precision agriculture, enhanced by artificial intelligence (AI), offers promising tools like remote sensing, intelligent irrigation, fertilization management, and crop simulation to boost agricultural efficiency and sustainability. Reinforcement learning (RL), in particular, has outperformed traditional approaches in optimizing yields and managing resources. Yet, widespread AI adoption remains limited by discrepancies between algorithmic recommendations and farmers’ practical experiences, local knowledge, and traditional practices. To bridge this gap, our study emphasizes Human-AI Interaction (HAII), specifically targeting transparency, usability, and trust in RL-driven farm management. We employ a well-established trust framework—consisting of ability, benevolence, and integrity—to construct a novel mathematical model quantifying farmers’ confidence in AI-based fertilization strategies. Farmer surveys conducted specifically for this research highlight critical misalignments, and these insights are incorporated into our trust model, subsequently integrated into a multi-objective RL framework. Unlike previous methods, our approach directly embeds trust into policy optimization, ensuring AI-generated recommendations are technically robust, economically feasible, context-sensitive, and socially acceptable. By aligning technical performance with human-centered trust, this research provides a practical path toward broader AI adoption in agriculture.