Some people have a knack for communicating with furry friends. Whether they can just understand dogs’ body language, distinguish between cats’ meows, or even pick up on the slightest difference in their snake’s mannerisms, there is no doubt that animals have ways of communicating, as long as humans are willing to listen. 

Beyond anecdotes, there is real evidence that humans are not alone in our ability to communicate. Many animals like elephants, ravens, dolphins, and parakeets, all show evidence of using names to refer to one another. Whales, which have brains six times larger than humans, use clicking sounds, or codas, similar to Morse code to talk to each other in the depths of the ocean. Analyses of whale recordings have found that their languages use complicated structures and organization not far off from our own. Other intelligent species like crows, famous for comprehending numbers and building tools, have astounded researchers with the complexity of their speech. Recursive language structures, such as sentences within sentences, were once thought to be unique to human language. In 2022, researchers made a linguistic breakthrough in finding the same structures in crow caws, and this discovery has since been found in monkeys and songbirds.

Now that we are aware of the complexity of animal language systems, we seek to understand them. Leading groups in this field, such as the Earth Species Project and Project Cetacean Translation Initiative bring together AI researchers, biologists, and conservationists to build machine-learning models that analyze extensive recordings of animal speech. These powerful models detect patterns and relationships among recordings of animal speech. This means they don’t need to create a rosetta stone that directly translates between individual words. Instead, they interpret language by building a map of word relationships, similar to how “mother” is to “daughter” as “father” is to “son.” They plan to build models similar to ChatGPT, which work by analyzing relationships between words in the English language.

Beyond natural fascination, there are real applications within conservation to understanding animal languages. For many species existing only in captivity, scientists struggle to recreate the conditions optimal for reproduction. This has led to diminishing population sizes and the threat of extinction. Furthermore, this technology can be applied to veterinary medicine, ecology, and general animal science. The ability to decode the noises and body language of the animals, or even perhaps communicate back and forth with them, might be the deciding factor in whether or not an entire species can survive. 

Because there remains much uncertainty surrounding the impacts of an animal translation system, they must be mindful of the ethical implications of this work. While the ability to better understand and communicate with animals contains much potential in the conservation sphere, it can also be used for more nefarious purposes. Some fear that poachers will use this technology to imitate animals and lure in vulnerable targets. Additionally, even a scientist with the purest of intentions may cause harm to an animal by using computer-generated language impersonating vocalizations the scientist cannot fully understand. With few ways to test if the AI has completely accurately mapped the animal’s language onto our own, scientists cannot be fully certain of the authenticity of the vocalizations. For this reason, many are hesitant to use these language models outside of animals being held in captivity, as it remains uncertain how this can impact an ecosystem on a larger scale

The ability to not only understand, but engage with animal communication will have sweeping effects on the fields of biology, conservation, and veterinary medicine. Rapid advancements in the overlap of linguistics and machine learning have the potential to make this a reality. While there remains hesitation surrounding the applications of the technology, researchers in the field aim to keep it in controlled environments. And if we’re lucky, it may only be a matter of time before man’s best friend becomes a true conversational companion.