Artificial intelligence is listening to orca calls in real time and helping to reduce their exposure to underwater noise. The effort is focused on an endangered orca subspecies in the Salish Sea, off the coasts of the northwestern U.S. and western Canada, reports Mongabay writer Abhishyant Kidangoor. The southern resident orcas (Orcinus orca ater), made up of just three pods, are one of the world’s most endangered marine mammal populations. There are an estimated 76 individuals remaining in the wild, as of December 2025. Vessel traffic and underwater noise are active threats to their survival because orcas use clicks and echolocation to hunt and locate their pod in the ocean. Constant noise from vessels makes it more difficult for them to communicate and navigate. Noise from a typical modern ship can raise underwater sound levels by 12 to 17 decibels, at frequencies lower than natural ambient noises, which can be extremely  disruptive for orcas. The decibel scale isn’t linear; a 10-dB increase means the sound intensity is 10 times stronger. Sound also travels faster and farther underwater than in air. Research shows that the odds of orcas catching prey decreases by 12.5% for every additional decibel of maximum noise. The AI-powered tool, OrcaHello, was developed during a 2019 hackathon event, and tracks the orcas’ movements by detecting their calls through underwater audio livestream. The team behind OrcaHello then trained a machine-learning model to recognize the calls of the specific orca subspecies so they can detect when the pods approach the port or…This article was originally published on Mongabay

From Conservation news via This RSS Feed.