Israeli researchers have developed an artificial intelligence system that detects when fish are hungry by analyzing the sounds they make while eating, a tool scientists say could make aquaculture more efficient and environmentally sustainable.
The system, created by the National Center for Mariculture in Eilat and the Israel Institute of Technology, uses machine‑learning algorithms trained to recognize the distinct clicking sounds sea bream produce when biting food pellets.
By filtering out background noise from fish farms, the technology can track feeding behavior in real time and determine whether fish are hungry, satiated or stressed.
Researchers said feeding control remains one of aquaculture’s biggest challenges. Farms typically rely on weight‑based estimates and feeding charts, which can lead to overfeeding — raising costs and polluting water — or underfeeding, which slows growth and increases disease risk.
The sound‑based approach, set to be published in the March issue of Computers and Electronics in Agriculture, offers a scalable and automated method to optimize feeding and improve fish welfare, the team said.
Early tests also show potential for detecting stress and aggression in shrimp and oysters, suggesting broader applications across aquaculture.
