Artificial intelligence in fisheries and aquaculture: Bridging algorithmic innovation with sustainable real-world deployment

Artificial intelligence (AI) is increasingly transforming fisheries and aquaculture by enabling data-driven monitoring, prediction, automation, and decision support across production, health management, resource governance, and value-chain systems. Despite rapid methodological advances in machine learning, deep learning, computer vision, Internet of Things (IoT), and digital infrastructures, the practical deployment of AI in aquatic food systems remains limited and uneven. This review critically evaluates recent developments in AI applications across fisheries genetics, aquaculture production, fisheries management, environmental monitoring, and enabling technologies, with specific emphasis on methodological robustness, validation practices, and real-world applicability. The synthesis reveals several systemic limitations: dominance of image-based models with limited multimodal integration, over-reliance on accuracy as a performance metric, insufficient external validation, and a predominance of proof-of-concept studies lacking operational scalability. Additional constraints include fragmented datasets, environmental variability, infrastructure limitations, and unresolved governance and ethical concerns. Rather than focusing solely on technical advancement, this review highlights the need for a paradigm shift toward robust, interpretable, and context-aware AI systems that are validated under field conditions. Future progress will depend on improved data standardization, multimodal integration, reproducibility, interdisciplinary collaboration, and equitable access for small-scale stakeholders. AI holds significant potential to enhance the efficiency and sustainability of aquatic food systems, but its long-term impact will depend on bridging the gap between algorithmic performance and real-world deployment.