AI decodes the calls of the wild - Nature

Marina: Welcome to Deep Voices — I'm so glad you're here. I'm Marina, a marine biologist who has spent more years than I can count completely enchanted by whales and dolphins, and this show is my love letter to them and to the science trying to understand their extraordinary minds. Marina: Before we dive in, a huge thank you to today's sponsors — Ada, who helps organizations integrate A I into their real-world workflows in ways that actually make sense; Ago Consulting — that's "Ahh-Go" — your expert guides through the world of silicon development, from A I all the way to SoC; and Zen Semiconductor, an umbrella company with a remarkable portfolio spanning the silicon industry. We couldn't do this without them. Marina: And today? Oh, today we have such a good episode. We're going to talk about a scientific effort to decode the secret language of a rarely-studied dolphin species. We're going to look at how A I is cracking open the whole field of animal communication. We're going to sit with something genuinely heartbreaking — a humpback whale mother and her stillborn calf. And we'll check in on some of the conservation and technology tools keeping cetaceans safer. Let's go. Marina: Okay, thread one — Voices of the Deep. And I want to start by introducing you to a dolphin you probably haven't met before. The rough-toothed dolphin. Steno bredanensis. Even the name sounds mysterious, right? These animals live in warm, deep, offshore waters — the open ocean, far from shore — which is actually a big part of why we know so little about them. They're not the bottlenose dolphins you see at the beach or in documentaries every other week. They're sleek, oddly distinctive-looking — they have this long, cone-shaped head that flows smoothly into the beak without the usual crease most dolphins have — and they tend to travel in groups, hunting together, living together. But their voices? Until recently, almost completely unstudied. Marina: A new study published in Nature has changed that. Researchers went out to observe free-ranging rough-toothed dolphins in the wild — which is already no small feat, given how far offshore these animals tend to be — and they recorded their whistles. Not just any whistles. They were specifically looking for repeated whistle types. Because here's the thing about dolphin communication that I find endlessly fascinating: dolphins don't just make random sounds. They produce structured, repeatable vocalizations. And the whistle is one of the most important tools in their acoustic toolkit. Marina: Now, you've probably heard of signature whistles — those unique, individual whistles that bottlenose dolphins develop early in life, essentially their names. They use them to identify themselves, to find family members, to maintain contact across distance. It's one of the most striking examples of individual identity in the animal world. But the rough-toothed dolphin? We didn't know what their whistle repertoire looked like. We didn't know if they had signature whistles, shared whistles, how complex or simple their communication might be. Marina: What the researchers found was that yes — rough-toothed dolphins do produce repeated, classifiable whistle types. They were able to identify and categorize these calls, pulling structure out of what might have seemed like acoustic noise. And that matters enormously. Because every time we document that another species has a structured, repeatable vocal system, we're adding a piece to this much larger puzzle about the evolution of communication, about how intelligence and sociality and vocal complexity are all tangled up together in the ocean. Marina: Rough-toothed dolphins are known to be socially complex — they form lasting bonds, they cooperate, they've even been observed in cross-species associations with other dolphins and whales. So it would make sense that their communication reflects that complexity. And now, for the first time, we're beginning to see it. We're beginning to hear it. That, to me, is what science is for. Marina: Thread two — Culture and Decoding. So let's talk about the bigger picture. Because what's happening right now in cetacean research — and in animal communication research more broadly — is genuinely historic. A I is changing everything. Marina: An article published in Nature lays it out beautifully: A I is being used to decode the calls of wild animals in ways that were simply impossible before. The sheer volume of acoustic data that researchers can now process — hours and hours of underwater recordings that would have taken human analysts years to go through — can now be analyzed by machine learning systems in a fraction of the time. And not just analyzed for the presence of sounds, but for structure, for pattern, for meaning. Marina: The most ambitious example of this is Project CETI — that stands for Cetacean Translation Initiative. It's a collaboration between marine biologists, linguists, cryptographers, and A I researchers, all focused on one audacious goal: understanding sperm whale communication. Sperm whales produce these click sequences called codas. Short, rhythmic bursts of sound that whales exchange during social interactions. And researchers have been recording them for decades. But the codas are complex, and the contexts in which they're produced are hard to study because sperm whales dive so deep, for so long. Marina: What Project CETI is doing is deploying an unprecedented recording infrastructure — underwater listening devices, tags on whales, even robotic fish to eavesdrop without disturbing the animals — and feeding all of that data into machine learning models. The hope is that by analyzing tens of thousands of codas in context — who made the sound, who responded, what happened next — the A I can begin to map out something like a grammar, a structure, a meaning. We're not there yet. But the fact that we're asking the question with the tools to potentially answer it? That's extraordinary. Marina: And it connects so beautifully to what I was saying about the rough-toothed dolphins. Every species we document, every whistle type we classify, every coda we record — it's all data. It's all pieces. The more we gather, the better our models become, and the closer we get to genuinely understanding what these animals are saying to each other. I sometimes think about what that moment will feel like — the moment when we can say, with scientific confidence, that we understand even a fragment of what a whale is communicating. I think it will change us. I really do. Marina: There's also something important I want to say about the cultural dimension of all this. Because cetacean communication isn't just individual — it's cultural. Whale songs, for instance, spread across ocean basins. A new song type emerges in one population of humpback whales, and within years it has traveled thousands of miles, picked up and spread by other whales who learned it from their neighbors. That's cultural transmission. That's learning. And dolphin populations have distinct dialects — groups of orcas, for instance, have call types that are specific to their family pod, passed down through generations. When we lose a population, we lose a dialect. We lose a culture. That's why this research isn't just scientifically fascinating — it's urgent. Marina: Thread three — A Cetacean and People Story. And I have two stories I want to bring you today, because I can't choose between them. One will break your heart a little. The other will give you hope. Marina: First, the one that broke mine. Smithsonian Magazine reported on something that researchers documented — a humpback whale mother who had given birth to a stillborn calf. And rather than leaving, rather than moving on, she stayed. She carried the calf. She touched it, repeatedly, with her fins. She looked at it. She stayed with her baby for an extended period, returning to it again and again. Marina: Now — the question the researchers were wrestling with, and that the article asks head-on — was she grieving? And that's a genuinely hard scientific question, because grief is an internal state, and we can only observe behavior. What we can say is that this behavior — staying with a dead offspring, making repeated physical contact, what researchers sometimes call "epimeletic behavior" or care-giving behavior toward a deceased individual — has been documented in multiple cetacean species. Orca mothers have been observed carrying dead calves for days. Dolphins have done the same. It's not a one-off anomaly. It's a pattern. Marina: What does that mean? It means these animals have social bonds deep enough that their loss produces a prolonged behavioral response. Whether we call it grief or not — whether that word applies — the bond was real. The attachment was real. And watching a humpback whale mother touch her stillborn calf and look into its eyes, whatever we name that, it is something. It deserves to be witnessed, and it deserves to be taken seriously. Marina: I think about this story in the context of everything else we've been talking about today — the complex communication, the cultural transmission, the A I trying to decode what whales are saying. Because the more we understand about the inner lives of these animals, the more the stakes of losing them become clear. North Atlantic right whales, for instance — one of the most critically endangered large whale species on the planet, with fewer than four hundred individuals remaining — have those same deep social bonds. They communicate. They have family relationships that last decades. And they are teetering on the edge. Marina: Which brings me to the hopeful story. The New England Aquarium has been at the forefront of right whale research and conservation for years, and they recently celebrated Right Whale Day — a moment to raise awareness about these magnificent, imperiled animals and the work being done to protect them. The Aquarium's right whale research team has contributed enormous amounts of data about right whale behavior, health, and population dynamics, including tagging and tracking individuals across their range. Marina: And then there's a story out of Stephen F. Austin State University, which is leading a drone development project called PHASM — that's the Photogrammetric Health Assessment of Small cetaceans with Multi-rotor drones. Researchers there are using specially designed drones to conduct health assessments of wild dolphins. The drones fly over dolphin groups and collect high-resolution aerial imagery, which researchers can then use to measure body condition — essentially, how healthy and well-fed individual animals are — without ever having to touch or disturb them. It's non-invasive, it's scalable, and it produces data we simply couldn't gather any other way. Marina: Think about what that means in practice. You can track whether a dolphin population is declining in health over time — whether animals are getting thinner, which can signal food scarcity or environmental stress. You can do it across large numbers of individuals. You can do it repeatedly. And combined with the acoustic and behavioral data that researchers like the ones studying rough-toothed dolphin whistles are gathering, you start to build a genuinely comprehensive picture of what's happening with these animals, in their world, on their terms. Marina: That integration of tools — drones, hydrophones, A I, long-term observation — that's the future of cetacean research. And it gives me real hope. Not naive hope, not the kind that ignores the scale of the threats. But the kind of hope that comes from watching smart, dedicated people bring everything they have to a problem that matters. Marina: Because these animals matter. The humpback mother who stayed with her calf — she matters. The rough-toothed dolphins trading whistles somewhere out in the deep, warm open ocean — they matter. The four hundred remaining right whales, each one known individually by researchers who have spent careers learning to recognize them — they matter. And the fact that we are finally, finally developing the tools to listen to them, to really listen — that feels like one of the most important things happening in science right now. Marina: So here's what I want you to take away from today. The ocean is full of voices. Complex, structured, culturally transmitted, emotionally charged voices. And we are just beginning to learn how to hear them. A I is helping. Drones are helping. Decades of patient fieldwork are helping. And every person who listens to a show like this, who cares, who pays attention — you're helping too. Marina: Deep Voices, brought to you by Ada, Ago Consulting, and Zen Semiconductor — three organizations investing in the technology that helps us understand our world better. We're grateful for their support. Marina: If today's episode moved you, please subscribe wherever you get your podcasts, and share it with someone who loves the ocean, or who needs to. Check the link in show notes for more on all the research and stories we talked about today — the rough-toothed dolphin study, Project CETI, the PHASM drone project, right whale conservation, and the humpback mother who stayed. Marina: Stay curious. Keep listening. The whales have been talking for millions of years. It's about time we learned to hear them.