AI technology could save Icelandic fishing industry billions

A new technology that uses artificial intelligence to predict the distribution of commercial fish stocks—such as cod, haddock, and redfish—could save the Icelandic fishing industry billions of ISK, according to Svanur Guðmundsson, a fisheries scientist and CEO of the Blue Economy.

The system is based on a neural network model trained on more than 1.2 million hours of fishing data collected since 2008 from Icelandic fishing companies. The data includes bottom temperature, depth, and catch per unit effort, and comes from all of the country’s largest fishing firms. According to Guðmundsson, this is the first time such comprehensive data has been used for predictive modeling of this kind.

“These data sets have never before been combined or analyzed in this way,” he says. All information is anonymized and not traceable to individual vessels, as it is subject to strict confidentiality.

By aggregating the data, the model can forecast fish stock movements well into the future. It also incorporates variables such as fishing location, weather conditions, sea temperature, and depth to improve the accuracy of its predictions.