1 Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, QC, H3G 1M8, Canada.
2 Kenneth Lee Research Inc, Halifax, Nova Scotia, B3H 4H4, Canada.
3 Department of Building, Civil and Environmental Engineering, Concordia University, Montreal, QC, H3G 1M8, Canada. Electronic address: chunjiang.an@concordia.ca.

Marine oil spills are one of the most severe anthropogenic threats to oceanic ecosystems, coastal communities, and global economic stability. While traditional monitoring and response approaches have played a foundational role in oil spill management, their effectiveness is often constrained by limited accuracy, slow response times, and high operational risks. Recent advancements in artificial intelligence (AI), particularly machine learning, computer vision, intelligent sensing, and robotics, have reshaped the landscape of oil spill detection, assessment, and emergency response. This review provides a comprehensive synthesis of AI-driven methodologies currently available for use across the full lifecycle of marine oil spill management. The contents examine AI-enabled risk prediction, failure forecasting, and toxicological and ecological impact assessments; AI applications in oil spill fate and transport modeling, such as physics-informed methods to ensure physical consistency, deep learning architectures for trajectory prediction, and uncertainty quantification techniques that enable probabilistic hazard assessments; integrated remote sensing systems, including autonomous robots; and intelligent manufacturing of remediation materials and the evaluation of AI-based decision support systems. This comprehensive overview of current developments and practical applications, aligned with stakeholder needs, identifies key challenges and provides recommendations for research on data availability, model generalization, interpretability, and system integration to advance AI-enabled, resilient, and environmentally responsible marine oil spill management practices.