
What happened
Denise Herzing of the Wild Dolphin Project told WIRED World Fair that AI is finally letting her team find patterns in dolphin audio and video, in work with Georgia Tech and Google DeepMind.
Why it matters
The team has used the research to attempt two-way communication with an underwater wearable computer that plays dolphin sounds, but telling mimicry apart from understanding remains hard, especially in the wild.
Summaries like this, every morning.
Herzing's interest in dolphin communication dates back to a 1985 trip to the Bahamas to study Atlantic spotted dolphins, where she heard their telltale clicks and whistles. The intervening four decades produced countless hours of video and audio footage, but not the means to give that material structure.
The current collaboration with Georgia Tech and Google DeepMind is aimed at finding patterns in that archive, not at full translation. Separately, the team has rigged an underwater wearable computer that plays dolphin sounds, using them to indicate items like a scarf and habitat like sargassum, a seaweed, to see how dolphins respond.
Herzing flagged a gap between mimicking and understanding the function of a sound, and suggested that working in a sanctuary or more controlled environment might help researchers tell the two apart. What dolphins actually say to each other remains a bit of a mystery, and she framed the meaning of the patterns in dolphin culture as an exploration of its own.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
IVRy started integrating アイブリー with Ikkyu's RESZAIKO so AI can take restaurant phone reservations around the c…

Ziva, an AI agent that runs inside the Godot game editor, generates up to three takes of each requested sound…

Unifa released the beta of Sukusuku Log on October 27

Using VoiceStudio with the VoxCPM2 model, a 15.3-second sample produced a clone whose normal-speech similarity…

Cartesia, an AWS Partner, detailed how its Sonic (text-to-speech) and Ink (speech-to-text) models, built on St…

Google DeepMind announced Genie 4 on 2026年10月27日, billed as its first real-time world model that natively gene…
