Applications of Artificial Intelligence in Fisheries: From Data to Decisions

Summary

This narrative review synthesises recent advances in artificial intelligence for aquatic resource management, spanning fisheries genetics, precision aquaculture, management systems, and sensing infrastructure. Whilst AI offers substantial potential for automating detection, optimising operations, and detecting illegal activities, the authors identify critical barriers to implementation including performance degradation across environmental transitions, insufficient standardised datasets for rare taxa, inadequate uncertainty quantification in management decisions, and structural inequalities in technology access among smallholder producers. The paper advocates for domain-robust algorithms, multi-region benchmarks, and privacy-preserving data partnerships to realise AI's potential in fisheries.

Regional applicability

The findings on AI implementation challenges and the need for standardised benchmarks are broadly applicable to United Kingdom fisheries management and aquaculture, though the paper's global scope means specific recommendations would require contextualisation to UK regulatory frameworks, data infrastructure, and industry characteristics including the relatively smaller scale of UK aquaculture compared to major producing nations.

Key measures

Species detection accuracy, feed optimisation efficiency, water quality prediction, illegal fishing detection capability, bycatch monitoring, protected species tracking, acoustic echogram analysis, domain shift performance, model uncertainty quantification

Outcomes reported

The paper reviews AI applications across fisheries genetics monitoring, precision aquaculture, fisheries management, and sensing infrastructure, identifying technological advances and implementation barriers. It examines automated species recognition, feed optimisation, water quality forecasting, illegal fishing detection, and ecosystem monitoring systems.

Topic tags