Google DeepMind has created an AI model called DolphinGemma that can help decode dolphin vocalizations, supporting research efforts to understand how dolphins communicate. The model was trained on data from the Wild Dolphin Project (WDP), a nonprofit studying Atlantic spotted dolphins and their behaviors.
DolphinGemma is built on Google’s Gemma series of models and can generate “dolphin-like” sound sequences. It is efficient enough to run on phones, making it accessible for researchers. This summer, WDP plans to use the Pixel 9 smartphone to create a platform that generates synthetic dolphin vocalizations and listens to real dolphin sounds for a matching response.
The upgrade from the Pixel 6 to the Pixel 9 will enable researchers to run AI models and template-matching algorithms simultaneously, according to Google.
Source: https://techcrunch.com/2025/04/14/googles-newest-ai-model-is-designed-to-help-study-dolphin-speech