When beluga whales communicate with one another, certain vocalization cues reveal their social structures. When meerkats face an approaching predator, they use a complicated set of alarm calls depending on how close and dangerous it is.
And in the skies, birds also communicate danger — along with things like mating status — to others in their flock.
Now, scientists are harnessing the power of machine learning to build a program that can decipher all these forms of animal communication, and more, by analyzing huge data sets.
“Everything we understand about animal signals so far comes from the geniuses that observed them,” says Katie Zacarian, cofounder and CEO of the Earth Species Project (ESP), a nonprofit organization working on creating this tool. “We know that it will dramatically accelerate the field of ethology and animal welfare.”
Read More: Can Animals Learn Language Like Humans Do?
Coding Animal Communication
Biologists all over the world have been decoding animal communication for decades now. It’s a time-intensive process, usually involving recording sounds or observing other signals that animals make to communicate — such as the “waggle dance” that honey bees conduct to tell one another where to find nectar.
Some biologists record hours of sound from animals, then pair this with observations about behavior. A lion may growl as a warning, for example, while a lizard may wave its hand or puff out a dewlap.
In some cases, these observations are helped by relatively new technology like biologgers: devices that, once fit onto animals, record everything from sound and acceleration to heartrate and more.
ESP launched in 2017, after advances in machine learning introduced new possibilities in decoding communication. The organization’s first peer-reviewed paper, published in Scientific Reports, described a tool that isolated once voice in a recording of multiple people speaking.
It’s a tricky task, known to scientists as the “cocktail party problem.” From there, the organization began pairing behavior observations with communication signals by analyzing the data stored in biologgers.
ChatGPT for Animals
Its mission isn’t necessarily to unlock the “language” that a specific species uses to communicate, but rather to develop a tool that biologists can use to recognize the signals animals make in certain situations.
In part, the tool could work something like a ChatGPT for animals, says Sara Keen, an acoustic engineer and neurobiologist at ESP. It’s basically an algorithm that can accurately predict what communication signal will come next based on an animal’s behavior — or vice versa.
Though the process may take some time, Keen points to the huge advances in speech recognition software in recent years: This kind of software was pretty bad 15 years ago, but now people rely on their Amazon Alexa for all kinds of things. The work on animal communication just needs to catch up, she says.
The project has “a huge potential for change,” adds Zacarian. “Our main mission is around unlocking this understanding of nature and transforming our relationship with the rest of nature in doing so.”
Read More: AI Might Help Us Decode Whale Language
Animal Conservation Applications
While impressive in and of itself, creating a baseline tool to understand all animal communication also has huge implications for conservation and human-wildlife interactions. The ESP team is already working with whale researchers on a project that analyzes the sound of belugas in the St. Lawrence River.
The biologists there have revealed that the whales are likely divided into several different social groups; and knowing that there may be distinct cultures out there has important conservation implications.
If one group is struggling more than the others, for example, it may warrant extra attention from wildlife managers.
Similarly, better understanding of non-human communication can reveal the ways that specific matriarchal herds of elephants pass on information. Knowing that some elephants talk in similar or dissimilar ways to one another could prove important for improving the success of reintroduction projects, Zacarian says.
But there may be a dark side to teaching humans how to better commune with animals.
Weighing Pros and Cons
Poachers, for example, are already using recorded playbacks of mating sounds to lure colorful songbirds like yellow cardinals into traps. It’s possible that a better understanding of animal communication will help bad agents to better exploit nature.
Zacarian says it’s an ongoing conversation that the organization has: “It’s imperative that we consider its use and try to mitigate bad consequences.”
On the other hand, the potential benefits for conservation remain huge — including everything from predicting mass whale beachings to revealing the impacts of offshore wind turbines.
And on a larger level, a better understanding of animal communication in general could inject more empathy and interest in conservation efforts, which in turn could help to reduce some of the problems that humans have caused to ecosystems around the world, Zacarian says.
“Humans,” she says, “need to dramatically shift the way they are relating with the rest of nature.”