More data, more problems? The challenge of integrating AI into fish farming.
At the North Atlantic Seafood Forum (NASF) on Tuesday, Erlend Torgnes, CEO of Optimeering Aqua, addressed the tension between traditional fish farming expertise and the rise of AI-driven decision-making.
His talk, titled “I’ve been doing this for 30 years – do you really think I need AI to tell me how to plan production?”, underscored the challenges of integrating AI into an industry where farmers rely heavily on intuition and hands-on experience.
“Farmers’ gut feeling is great, but our programming can help prove it,” Torgnes said.
The AI Learning Curve in Aquaculture
Optimeering Aqua’s Bioplan decision support tool aims to enhance—not replace—farmers’ expertise by synthesizing data from more than 10 different sources monitoring cage conditions. However, introducing AI models into farming has led to friction, as farmers are often skeptical of technology telling them what they already believe they know.
“Farmers don’t know what they want, but they do know what good looks like,” Torgnes noted, emphasizing that AI should validate and refine existing knowledge rather than dictate decisions.
Key Lessons from AI Adoption in Fish Farming
Torgnes outlined five key takeaways from deploying AI in real-world farming environments:
- Data supports instinct – AI models should reinforce farmers’ expertise, not override it.
- Too much data is overwhelming – AI must provide actionable insights, not just raw numbers.
- Farmers need proof – AI recommendations must be demonstrably effective to gain trust.
- Iterative development is key – AI tools must evolve alongside farming realities.
- Decision support, not decision control – Farmers must remain in control, using AI as a refinement tool.
As AI continues to make inroads into aquaculture, Optimeering Aqua argues that its success hinges on bridging the gap between technology and traditional knowledge. The key challenge lies in making AI useful and intuitive rather than an imposed system.
