Emerging AI Solutions for Hazardous PET Waste in Marine Environments: A Review of Underexplored Paradigms

Yara Hossam, Hajar Nagdy, Rana Adel, Sherif Mehanny, Irene S. Fahim, Ola Gomaa, Tajalli Keshavarz, Ahmad Al-Kabbany

Polyethylene terephthalate (PET) pollution, due to its persistence, chemical recalcitrance, and widespread usage, represents a growing hazard to marine ecosystems. Its accumulation contributes to long-term ecotoxicological risks, food chain contamination, and environmental degradation. Addressing this challenge necessitates the adoption of scalable, efficient, and intelligent strategies for monitoring, collection, and degradation. While artificial intelligence (AI) has shown promise—particularly through machine learning (ML), deep learning (DL), and computer vision (CV)—these approaches have already been extensively covered in recent reviews and are therefore beyond the scope of this research.

Instead, this article focuses on underexplored and emerging AI technologies with untapped potential in marine PET management. These include reinforcement learning, generative AI, Edge AI, soft robotics, federated learning, explainable AI, decision support systems, and large multimodal models (LMMs). These technologies offer new capabilities for real-time decision-making, distributed data processing, autonomous biodegradation, and intelligent system design—critical tools in managing hazardous waste scenarios—yet remain largely absent from current PET-focused research.

We synthesize existing literature where applicable—such as in reinforcement learning, Edge AI, and soft robotics—while also investigating areas where research is still sparse or speculative, such as federated learning and LMMs. This addresses the need for a comprehensive review that not only maps current applications but also identifies conceptual gaps where emerging AI technologies can drive future innovation.

Through a bibliometric analysis of leading scientific databases and a thematic synthesis of technological trajectories, this article straddles the line between a traditional review and a forward-looking conceptual overview. It aims to inform researchers, inspire interdisciplinary collaboration, and advance sustainable, hazard-mitigating solutions for marine PET management in a rapidly evolving technological landscape.