Is the Pixel really translating dolphins? This is what’s happening with Google’s AI

Google has taken a bold leap into marine biology with an artificial intelligence system designed to decode dolphin communication. The project, known as DolphinGemma, is being tested on smartphones—yes, even your Pixel could soon help bridge the gap between humans and dolphins.

How Google’s AI is helping us understand dolphins

For over four decades, researchers at the Wild Dolphin Project (WDP) have recorded dolphin vocalizations, trying to decipher the meaning behind their clicks, whistles, and buzzes. Now, with Google’s help, this massive dataset is feeding DolphinGemma, a large language model adapted for marine acoustics.

By using audio processing tools like SoundStream, Google’s system identifies patterns in dolphin sounds, predicting what might come next in a vocal sequence. This makes it possible not only to understand dolphins better but also to create synthetic sounds for a two-way communication experience.

From phones to the ocean: Dolphin translation in real time

In a surprising twist, these complex tasks are being run on Google Pixel 6 smartphones, thanks to their real-time, high-fidelity audio analysis capabilities. Their portability, low energy use, and compact design make them ideal for underwater use with systems like CHAT (Cetacean Hearing Augmentation Telemetry).

Looking ahead, Pixel 9 devices will take over, bringing improved microphones, speakers, and processing power to deepen this cross-species dialogue. DolphinGemma may one day do more than interpret dolphin sounds—it could pave the way for communication with other intelligent species.

Google’s AI isn’t just enhancing your photos anymore—it might be helping you talk to dolphins.