AI-Driven Approaches for Detecting, Assessing, and Managing Microplastic Pollution in Agroecosystems

Microplastics are emerging contaminants in agricultural systems. They are sourced through plastic mulches, reclaimed wastewater irrigation, manure application, agrochemical packaging, and atmospheric deposition. Their movement through the soil–water–crop systems can degrade soil function, reduce yields, and contaminate food. In soil, microplastics accumulate, interact with biota, and may transfer to crops. Conventional microscopy and spectroscopy are labor-intensive and poorly suited for frequent monitoring of microplastics. This review investigates how artificial intelligence (AI) can enable scalable, data-driven assessment and mitigation. We synthesize machine learning and deep learning advances in computer vision for microscopy, AI-assisted FTIR and Raman spectroscopy, and predictive models linking particle size, shape, polymer type, and surface chemistry to ecological and toxicological outcomes. The review indicates that AI workflows often achieve 85–90% identification accuracy while reducing analysis time from hours to seconds. AI models enhance estimates of transport, ingestion, and bioaccumulation. It also helps in identifying contamination hotspots across agroecosystems. Key challenges include data availability for model testing and training, as well as a lack of interpretable models that align with regulatory and precision-agriculture needs within sustainable, precision agriculture and land management.