Artificial intelligence for microplastic pollution monitoring, predictive modeling, and risk assessment: advances, challenges, and future perspectives

Abstract

Abstract Microplastic (MP) pollution constitutes a pervasive and complex environmental crisis, with particles infiltrating ecosystems globally and posing multifaceted risks to ecological integrity and human health. The inherent heterogeneity of MPs (in size, shape, polymer type, and degree of weathering) coupled with their presence in complex matrices, presents formidable analytical challenges that constrain traditional detection, quantification, and source-tracking methods. This review synthesizes the transformative potential of Artificial Intelligence (AI) and machine learning (ML) in overcoming these bottlenecks and advancing every facet of MP research. This review details how AI-driven computer vision automates the morphological analysis of MPs, while ML algorithms enhance the speed and accuracy of spectroscopic polymer identification. Beyond analytics, AI facilitates predictive modeling of MP fate and transport, enables intelligent source apportionment, and optimizes emerging remediation technologies. Using the United States as a critical case study, we examine the interplay between a fragmented regulatory landscape and AI innovation. However, significant challenges persist, including data scarcity, algorithmic limitations, and the gap between laboratory validation and field deployment. The reviewed findings argue that the integration of AI represents a necessary paradigm shift, moving from descriptive observation to predictive and prescriptive science. For this potential to be fully realized, concerted efforts in standardized data generation, interdisciplinary collaboration, and the development of transparent, robust AI frameworks are urgently required to inform effective mitigation policies and remediation strategies on a global scale.