© Getty Images

What are they saying? -

Can animals communicate in ways we can understand? Can we bridge the species divide and not only understand their communication but also talk directly to them? These are questions that have fascinated humankind for generations.

© Shutterstock

Ancient connection -

While many Indigenous cultures believe animals can communicate intentionally, Western science has often avoided research suggesting human-like qualities in animals, fearing accusations of anthropomorphism—the attribution of human characteristics to non-human entities.

© Getty Images

First attempts -

In the '70s, scientists tried to teach human language to primates like Koko, a four-and-a-half-year-old gorilla, who gained fame for her ability to use sign language.

© Getty Images

Chatty friend -

Her instructor, Francine "Penny" Patterson (left), successfully taught Koko sign language, allowing her to communicate basic ideas and emotions. Patterson documented that the gorilla understood some 2,000 words of spoken English.

© Getty Images

Animal sound studies -

Bioacoustics, the study of animal sound, emerged in the 1920s. While marine mammal sounds are now familiar, the first recordings were made in the late 1940s.

© Getty Images

Field evolution -

Research on how dolphins use sound to "see" underwater and how different species use unique sounds grew rapidly in the 1950s and 1960s.

© Getty Images

On the record -

More recently, the field of digital bioacoustics has been using technology to study the sounds animals make. By recording and analyzing these sounds, scientists can learn about how animals communicate, move, and interact with their environment. 

© Getty Images

Too many sounds -

As recording animal sounds becomes more accessible, human scientists struggle to keep up with the sheer amount of recordings from different species. 

© Getty Images

AI use -

AI, however, thrives on large datasets. Just like language models learn from vast amounts of text, AI can learn to recognize patterns in animal sounds.

© Shutterstock

Decoding animals -

But while language models like ChatGPT learn from human language, how can AI translate animal communication? This requires a different approach, similar to how other AI software can generate images from text.

© Shutterstock

Multiple elements -

This "multimodal" analysis, understanding different forms of information like sounds and images, could be key to deciphering animal language.

© Getty Images

Not just sound -

Just as we use body language and gestures, animals also use physical cues alongside sounds. These actions can occur before, during, or after vocalizations.

© Shutterstock

Pattern hunting -

Researchers have cataloged these behaviors in a list called an ethogram. By training machine learning models to analyze this extensive list, they can uncover patterns in the data and gain a deeper understanding of complex animal communication.

© Shutterstock

Breakthrough -

Karen Bakker, a professor at the University of British Columbia and an expert in the field, believes that combining digital listening with AI could lead to two significant discoveries: the existence of language in non-human species and the possibility of interspecies communication.

© Getty Images

Google Translate for animal speech -

AI is already being used to decode animal communication. Neuroscientist Kevin Coffey helped develop DeepSqueak, a machine-learning tool that decodes rodent vocalizations, offering insights into its behavior.

© Shutterstock

Happy and sad? -

According to Coffey, rats use high-pitched "laughter" calls in positive situations, such as courtship and play, and lower-pitched calls in negative ones, like when they're ill or in pain. Coffey uses these sounds to monitor the well-being of his lab rats.

© Shutterstock

Deep dive -

Other groups such as the Project Cetacean Translation Initiative (CETI), led by whale biologist Shane Gero, are focusing on trying to understand a particular species, in this case, the sperm whale.

© Getty Images

Whale lingo -

For nearly 20 years, Gero recorded sperm whale vocalizations and their associated behaviors. He discovered that whales use specific sound patterns called codas to identify each other.

© Getty Images

Mimicking mom -

The study found whales learn these codas much like toddlers learn words and names: by imitating the sounds of adults around them.

© Getty Images

Positive results -

After manually decoding some of these codas, Gero's team explored using AI to accelerate the process. The algorithm successfully identified a small subset of individual whales from their codas with 99% accuracy.

© Getty Images

Learning to speak "whale" -

Project CETI also plans to deploy underwater microphones to continuously record whale vocalizations, aiming to train AI to understand and even "speak" whale language.

© Getty Images

Sending messages -

AI software has already made progress in both translating animal communication and transmitting information to animals. It can generate specific sound patterns to communicate with bat colonies or beehives.

© Getty Images

Busy bees -

Tim Landgraf, a researcher at Freie Universität Berlin, discovered that honeybees communicate through both body language and sound. He is now using deep-learning algorithms to track individual bees and analyze the impact of their communication on others.

© Getty Images

Bee signs -

Landraf identified sounds like "toots," "quacks," a "hush" or "stop" signal, a "whooping" danger signal, and various piping, begging, and shaking signals. These signals direct both collective and individual bee behavior.

© Getty Images

Chat robot -

Landgraf also created RoboBee, a robot that can enter a hive and communicate with bees. It can give instructions and even perform the waggle dance, a complex communication pattern used to indicate nectar sources.

© Getty Images

Understanding pets -

Machine learning could also help us better understand our pets. Animal behaviorist and conservation biologist Con Slobodchikoff is developing an AI model to translate dogs' facial expressions and barks for their owners.

© Shutterstock

Body language - According to Slobodchikoff, pets often communicate using multiple signals, such as barking and body language. However, we tend to focus solely on sound, missing crucial visual cues.

© Shutterstock

Nature's SOS -

Decoding animal sounds and transmitting information to animals could revolutionize conservation efforts. Researchers can use AI to track endangered species, identify behaviors contributing to their decline, and aid in their protection, like guiding honeybees to safe nectar sources.

© Getty Images

Dangerous risks -

On the other hand, scientists also worry that this technology could be misused for harmful purposes, such as luring endangered species into traps or manipulating their behavior.

© Getty Images

Philosophical implication -

Some scientists compare this breakthrough to the invention of the telescope, which shifted our perspective on Earth's place in the universe. AI-powered research into animal communication could similarly reshape our understanding of our relationship with other species.

Sources: (Scientific American) (Deutsche Welle) (Science Direct) (Tech Times) (BBC) (Gray DI)

See also: It turns out animals understand a lot about death

© Shutterstock