Humans have long viewed the world through an anthropocentric lens, limiting our understanding of other species. Despite scientific efforts, language remains a significant barrier. However, recent advances in generative AI offer hope. Projects like the Earth Species Project and Project Ceti aim to decode non-human communication, using AI to analyze animal sounds and link them to behavior and emotions. This breakthrough could revolutionize conservation and deepen our connection to the animal kingdom.
Non-profit Earth Species Project seeks to “decode non-human communication” using AI. Its roadmap involves gathering samples of animal “speech”, and using AI to search for patterns to link vocalizations with behavior (e.g. when a specific bird call leads other birds to look for shelter). In addition, it may be used to link sounds with emotional states, letting us know when an animal call may be a cry for help, for example.
Project Ceti aspires to “listen to and translate the communication of sperm whales”. Their process includes four phases: firstly, monitoring whale sounds using hi-tech equipment (underwater microphones, aerial drones, and swimming robots). They then process the raw data before beginning to train the CETI technology to link behavior with language. Finally, they plan to validate the model with scientists specializing in linguistics, phonology, and morphology. This is all made possible through machine learning, which can process vast amounts of data, pulling out intricate patterns and insights.
Marine biologist David Gruber introduces Project CETI (Video, 6.5 min.)