Project CETI uses AI to decode sperm whale clicks and finds vowel-like sounds

A man who listens to whales

Marine biologist David Gruber founded Project CETI to decode sperm whale communication using large language models and machine learning. The team has found vowel-like features in sperm whale codas, regional dialects learned through matrilineal clans, and a babbling phase in calves. Working with linguists and cryptographers, they built WhAM, a synthetic model of whale language, while developing ethics guidelines for interspecies communication.