Decoding animal communication with AI

Researchers are using AI to investigate how animals communicate, with the goal of better understanding other species’ inner lives. Advances in recording technology, from ocean-floor hydrophones to wearable biologgers, have gathered detailed data on animal sounds. AI can help distinguish calls, identify the animals producing them, and find patterns.

The Earth Species Project, a nonprofit founded in 2018, recently used machine learning to detect carrion crow vocalizations and determine which bird made each call. Its NatureLM model, trained on human speech, music, and animal sounds, is intended to interpret vocalizations across species. The prospect of translating animal communication, however, remains theoretical.

Understanding meaning requires linking sounds to behavior and surrounding events, a painstaking task for species that are difficult to observe. Animals also communicate through visual cues, touch, and pheromones that sound-focused systems may miss. Tel Aviv University neuroecologist Yossi Yovel warns that models could confuse acoustically similar sounds with different meanings. Without a reliable way to verify translations, accurate interpretation remains difficult, and whether humans will achieve two-way conversations with animals is uncertain.