Using data science to detect and prevent fish suffering in real time
Traditional aquaculture welfare assessment relies on periodic human observation — insufficient given the scale, underwater environment, and subtle behavioral indicators of fish welfare. Emerging technologies including computer vision, acoustic monitoring, water quality sensors, and AI-based behavioral analysis are enabling continuous, objective welfare monitoring that was previously impossible. These technologies have potential to transform fish welfare from a periodic concern to a continuously managed dimension of aquaculture operations.
Underwater camera systems combined with AI image recognition can monitor:
Continuous water quality monitoring (dissolved oxygen, temperature, pH, ammonia) enables immediate response to conditions that compromise welfare. Smart feeding systems that detect appetite changes (fish not rising to feed) can identify welfare problems 24-48 hours before clinical disease. Acoustic monitoring can detect gill cleaning behavior and erratic movement associated with lice burden. Integration of these data streams with AI decision support allows earlier intervention and better welfare outcomes.
The Operational Welfare Indicators (OWIs) for Atlantic salmon, developed by Norwegian research institutions, provide a validated framework for welfare scoring in commercial farms. OWIs assess wound prevalence, fin damage, cataract levels, and behavioral patterns. Digital welfare scoring platforms enable farm comparison and longitudinal tracking. These tools are moving welfare from intuitive assessment to evidence-based management.