Emerging Artificial Intelligence Applications in Capture Fisheries Monitoring and Assessment: A Bibliometric and Synthesis Review

Artificial intelligence (AI) is increasingly used to monitor and assess capture fisheries, addressing persistent limitations of conventional observation systems that are labor‐intensive, costly, and difficult to scale. Despite rapid technological expansion, AI development in this field remains fragmented, limiting a comprehensive understanding of its structure and direction. This study aimed to map the development, intellectual structure, and thematic evolution of AI research in capture fisheries monitoring and assessment through a bibliometric and synthesis‐based review of Scopus‐indexed literature (2015–2026). Following a documented, large language model‐assisted screening protocol with human verification, a final corpus of 495 journal articles was analyzed using an integrated bibliometric and science‐mapping workflow. Results showed a marked, near‐exponential increase in publication output after 2020. Rather than confirming a pre‐defined three‐domain structure, keyword co‐occurrence analysis revealed that the field's thematic organization follows technical task type (detection, classification, and operational/vessel management) more closely than application domain, with acoustic AI failing to form an independent thematic cluster and representing the smallest share of the corpus (8.1%), behind visual AI (44.0%) and vessel monitoring (28.9%). For synthesis, we nonetheless compare the field across three functional categories visual AI for fish detection and measurement, vessel monitoring using AIS/VMS data, and acoustic survey support finding that the field is methodologically converging around a shared AI/ML toolkit while remaining largely siloed at the technical‐application level. Although research productivity remains heavily concentrated in China (85.9% of records) and international collaboration is limited (23.64%), the field is moving beyond isolated proof‐of‐concept studies toward more operational, management‐relevant applications. Overall, this study provides an evidence‐grounded, rather than assumed, understanding of how AI research in capture fisheries is thematically structured, and identifies acoustic AI and cross‐domain integration as priority areas for future development toward fisheries‐oriented, evidence‐based management.