Can Artificial Intelligence Decode the Language of Animals?

Scientists are utilizing artificial intelligence and machine learning technologies to understand the communication methods of many species, ranging from birds to elephants and whales.

The US-based non-profit Earth Species Project is developing artificial intelligence models to analyze animal sounds and identify communication patterns. The institution's developed NatureLM-audio model can examine large audio archives by recognizing and classifying animal sounds. In tests conducted on recordings from Australia's FrogID project, this model achieved a 99 percent accuracy rate in distinguishing frog sounds and an 82 percent accuracy rate in identifying specific frog species.

The Earth Species Project is conducting collaborative studies with researchers examining species such as killer whales, crows, and finches. The ZF-AIM model, developed with McGill University, generates real-time artificial intelligence-based bird sounds to examine the vocal interactions of finch birds. Meanwhile, Google DeepMind has developed the DolphinGemma model, trained on Atlantic spotted dolphin sounds collected by the Wild Dolphin Project, to analyze patterns in dolphin vocalizations.

Project CETI, an international initiative, uses artificial intelligence, robotics, and underwater recording technology to examine the communication of sperm whales. A study published in the journal Nature Communications in 2024 examined over 8,700 sperm whale codas, revealing elements of rhythm and tempo. However, researchers emphasize that detecting sound patterns does not mean knowing precisely what these sounds mean, and that sounds must be evaluated alongside behaviors and environmental conditions.