# Interspecies Communication This note brings together experiences and discussions related to efforts to understand and facilitate communication between human and non-human species. It includes research initiatives, AI applications and conceptual frameworks that explore how meaning can be shared across species boundaries. ## Entries **14/04/2025**: [DolphinGemma](https://blog.google/technology/ai/dolphingemma/) is a foundational AI model developed by Google and deployed by the [Wild Dolphin Project](https://www.wilddolphinproject.org/). Trained on decades of dolphin vocalizations, it can recognize, predict, and generate dolphin-like sound patterns. It's also integrated into CHAT, a system for two-way interaction using synthetic whistles. **10/12/2024**: The Cetacean Translation Initiative ([Project CETI](https://www.projectceti.org/)) is an interdisciplinary effort, funded by David Gruber, that aims to decode sperm whale communication through the use of machine learning, extensive hydrophone arrays, and behavioral context tagging. The project seeks not only to expand the frontiers of interspecies understanding but also to [contribute to marine conservation](https://www.nature.com/immersive/d41586-024-04050-5/index.html). **01/08/2023**: [Earth Species Project](https://www.earthspecies.org/) (ESP) is a research initiative aiming to decode non-human communication using machine learning. Led by [Aza Raskin](https://www.youtube.com/watch?v=3tUXbbbMhvk), the project uses AI to detect patterns in animal sounds, starting with species like whales and primates. ESP models turn semantic meaning into geometric space, allowing patterns of meaning to be compared across human languages — and potentially across species.