AI is moving forward with plans to help humans communicate with animals, and a promising step comes from a new tool called DolphinGemma.
Developed by Google DeepMind, DolphinGemma is a large language model (LLM) trained on dolphin audio data to help decode how dolphins talk to each other. The goal? To someday allow human-animal communication, a concept once thought to be purely science fiction.
“I like to think that we will be able to talk to animals at some point,” said Drew Purves, Nature Lead at Google DeepMind, on the company’s podcast.
What Is DolphinGemma?
DolphinGemma is DeepMind’s pioneering AI model designed to analyze and interpret dolphin vocalizations. Built in collaboration with Georgia Tech and the Wild Dolphin Project, the model separates, tokenizes, and processes dolphin sounds into structured data.
This method mirrors how LLMs like ChatGPT are trained on human language—only this time, the language belongs to dolphins.
