Can AI Decode Animal Communication?

There are other hurdles as well. Many animals communicate not only through sound, but also through visual cues, touch, pheromones, or a combination of signals—some of which are difficult for humans to perceive. Without multiple species serving as a kind of Rosetta Stone, it is also difficult to know whether a translation is accurate.

How technology is advancing animal communication research

Over the past 50 years, the proliferation of recording and tracking devices has made it easier to observe the sophisticated ways animals communicate. Hydrophones can capture humpback whale sounds at the bottom of the ocean, while small wearable devices called biologgers can record every sound made by a bat.

Advances in artificial intelligence over the past decade have further improved researchers’ ability to analyze animal communication. AI can detect different calls, distinguish which sounds come from which animals, and identify patterns in large datasets. Researchers at the Earth Species Project (ESP), a nonprofit founded in 2018 to decipher animal communication, recently used machine learning to detect the calls of carrion crows and determine which bird made each call.

Researchers believe more advanced AI could eventually translate the chatter of living things. Large language models such as Claude and ChatGPT are effective at identifying relationships between words, explains Sarah Keene, one of ESP’s behavioral ecologists. ESP hopes its NatureLM model, trained on human speech, music, and animal sounds, will be able to interpret vocalizations from a variety of species.

Why translating animal sounds remains difficult

However, this promise is still theoretical. Yossi Yobel, a neuroecologist at Tel Aviv University, worries that AI models may overemphasize acoustic similarities and lump similar sounds with different meanings together.

What about two-way communication? It is not clear when—or whether—that will happen. Still, “many animals have similar enough evolutionary structures that we should be able to understand what they’re saying,” Slobochkov says.

Progress meter: 4/10

Researchers have amassed a wealth of data on animal communication, but they are still figuring out how to decipher it all.