Studying the Impact of Explainable AI in Digital Agriculture Solutions

Although agriculture has traditionally appeared to be a perpetual industry, it has encountered a gradually rising set of significant challenges in recent years, including the alarming depletion of natural resources, the rising water and food demands, and the limited amount of arable land. In response to these challenges, the agricultural sector has seen an increased effort to embrace the digital revolution, harnessing emerging technologies to optimize sustainable agricultural processes and provide valuable support for informed decision-making. Given the persistent lack of transparency in machine learning models, farmers' perceived complexity and low value of smart agriculture solutions, this research focuses on studying the impact of explainable AI techniques and multimodal data on farmers' user experience in digital agriculture solutions. In this context, we propose a novel collaborative platform, AgriUXE, particularly tailored for AI-driven digital agriculture applications. The platform will focus on augmenting the explainability of both captured multimodal data and machine learning models' predictions. Moreover, it is crucial to evaluate how an optimized user experience, achieved through the development of transparent data-driven solutions in collaboration with key farm stakeholders, influences the expectations of small and medium-sized farmers regarding smart farming technologies.