Observe's AI registers salmon activity rising with gill disease

AI system

System: Observe Technologies behavioral monitoring algorithm (patented convolutional neural network, EP3644717)
Model: Not assessed
Release: Not assessed
Application developer: Observe Technologies, Richmond, UK
Role: Not assessed
Configuration: Not assessed
Inputs and outputs: Not assessed
Autonomy: Not assessed
Workflow: Feed cameras in all cages plus 4–5 extra cameras in one cage per farm. Weekly gill scoring, welfare indicators, mortality and feeding-rate records. A summer 2023 gill-health episode at both farms.
Human oversight: Not assessed
Relation to AI: The vendor's neural network quantified a behavioral response that coincided with gill disease already identified by weekly scoring. The study shows correlation, not an alert that changed what happened. Vendor staff co-authored the peer-reviewed paper.
AI contribution: The vendor's neural network quantified a behavioral response that coincided with gill disease already identified by weekly scoring. The study shows correlation, not an alert that changed what happened. Vendor staff co-authored the peer-reviewed paper.

Outcome identifiers: Not assessed
Benefit: fish activity rose from 25.6% to 43.6% (Farm A) and 24.9% to 32.6% (Farm B)
Category: Improved Knowledge Discovery
Status: Not assessed
Beneficiary groups: Organizations. People.
Beneficiaries: Not assessed
Reach: Not assessed
Distribution and access: Not assessed
Duration: Not assessed
Expected?: Yes
Costs and harms: Not assessed

Studies: Not assessed
Study design: Not assessed
Measurement: Not assessed
Setting and sample: Not assessed
Measures: fish activity rose from 25.6% to 43.6% (Farm A) and 24.9% to 32.6% (Farm B)
Comparison: Not assessed
Results: fish activity rose from 25.6% to 43.6% (Farm A) and 24.9% to 32.6% (Farm B)
Additional value from AI: Not assessed
Limitations: Not assessed
Denominators: Not assessed
Materials for replication: Not assessed

Participants: Developer: Observe Technologies, Richmond, UK. Operator: Two commercial Atlantic salmon marine farms in Scotland, studied by the University of Stirling.
Jurisdictions: Developer: United Kingdom. Operator: United Kingdom.
Ownership and control: Not assessed
Dependencies: Not assessed
Cross-border connections: Not assessed
Geopolitical implications: Not assessed

Conditions for replication: Gill disease severe enough to alter shoaling and distribution. Camera coverage of the cage volume.
Replication attempts: Not assessed
Later findings: No later report on the Observe algorithm's use on these or other farms after the 2023 season was found.
Availability: Not assessed
Institutional decision: Not assessed
Grounds: Not assessed
Action and result: Not assessed
Success criterion: Not assessed

Assessed benefits: Organizations. People.
Depth: Not assessed
Reasons: Organizations: Gill disease was scored weekly by conventional means. The paper does not report any farm decision changed by the activity signal. People: No outcome for farm staff or fish welfare beyond the behavioral correlation is documented.
Observed or intended: Not assessed
Rubric: Not assessed
Assessor and date: Not assessed
Alternative assessments: Not assessed
Fill: Independently studied

Kind: Not assessed
Related entries: Not assessed
Field: Aquaculture / fish welfare. Atlantic salmon sea cages.
Purpose: Extract group behavior (abundance, speed, shoal cohesion) from feed-camera video as a non-invasive indicator of fish health and welfare.
Dates: Date basis unassessed: 2023 (precision unassessed). Updated: 21 October 2026.
Stage: Not assessed
Version: SHA-256: f28425452cbe…

Peer-reviewed study Independent author

Burke M, Nikolic D, Fabry P, Rishi H, Telfer T, Rey Planellas S. Precision farming in aquaculture: assessing gill health in Atlantic salmon (Salmo salar) using a non-invasive, AI-driven behavioural monitoring approach in commercial farms. Aquaculture Science and Management 2:15 (2025). doi:10.1186/s44365-025-00020-8