- Google and Wild Dolphin project has developed an AI model trained to understand dolphin vocalizations
- Dolphingemma can run directly on pixel -smartphones
- It will be open sourced this summer
In most of human history, our relationship with dolphins has been a one -sided conversation: We speak, they squeak, and we nod as if we understand each other before we throw them a fish. But now Google has a plan to use AI to bridge this gap. In collaboration with Georgia Tech and Wild Dolphin Project (WDP), Google has created Dolphingemma, a new AI model trained to understand and even generate dolphin chatter.
WDP has collected data on a particular group of wild Atlantic spotted dolphins since 1985. The Bahamas-based pod has provided huge amounts of sound, video and behavior notes as scientists have observed them, documented any squawk and humming and trying to share what it all means. This treasure chest of sound is now being brought into dolphingemma, based on Google’s open Gemma family of models. Dolphingemma takes dolphin sounds as input, treats them using audio tokenizers like soundstream and predicts what vocalization can come next. Imagine AutoCompleTe, but for dolphins.
The model is very slim and can run on a Google Pixel. WDP already implements Dolphingemma in the field this summer using Pixel 9S in waterproof rig. These models will listen to, identify vocal patterns and help scientists mark meaningful sequences in real time.
Look at
Flipper speaks
But the ultimate goal here is not just passive listening. WDP and Georgia Tech are also working on a system called Chat (abbreviation for Cetacean Hearing Augmentation Telemetry), which is essentially a two -way communication system for humans and dolphins. Chat lets scientists assign synthetic flutes to objects Dolphins such as, including seagrass and liquid scarves, and then waits to see if the dolphins mimic these sounds to request them. It’s a bit like inventing a shared language, except with underwater microphones instead of flashcards.
Dolphingemma not only analyzes dolphin sounds after the fact; It helps to predict which sounds may come, enabling faster response times and smoother interactions. In essence, it is as a predictable keyboard for dolphins. The entire project is still at an early stage, but Google plans to open source dolphingemma later this year to speed up progress.
The initial model is trained on the vocalizations of the Atlantic Spotted Dolphins, but it can be theoretically adapted to other species with some tuning. The idea is to give other scientists the keys to AI so that they can apply it to their own acoustic data sets. Of course, this is still far from chatting with dolphins about philosophy or their favorite snacks. There is no guarantee that dolphin vocalizations map nicely to human -like language. But Dolphingemma will help to silence through many years of sound to meaningful patterns.
Dolphins are not the only animals that humans can use AI to communicate with. Another group of scientists developed an AI algorithm to decode the pig’s feelings based on their grinning, screams and snuff to help farmers understand their emotional and physical health. Dogfins, however, are undeniably more charismatic. Who knows, maybe one day you will be able to ask a dolphin for directions while you are sailing, at least if you don’t let go of your phone in the water.
