Scientists are cracking the code of interspecies communication. Is that a good thing?
Researchers led by Nicolas Mathevon used machine learning to decode the vocalizations of African striped mice in the Karoo, revealing distinct calls for neighbors versus strangers, challenging assumptions about animal communication complexity. The study, one of four finalists for the Dolittle Prize, demonstrates how AI can uncover hidden patterns in wildlife sounds, with potential future applications in deciphering dynamic information like stress levels.
Scientists studying African striped mice in South Africa’s Karoo semidesert have decoded their high-frequency squeaks, revealing nuanced communication patterns. Using 23 microphones over 12 days, researchers recorded 122,619 vocalizations from dozens of mice, identifying at least seven distinct squeak types. These calls signal identity—mice react strongly to unfamiliar voices, even fleeing when hearing a stranger’s call, while ignoring familiar nestmates. The team, led by Nicolas Mathevon of the University of Saint-Etienne in France, trained an artificial neural network—similar to those used in large language models—to analyze the data. The AI uncovered unique vocal signatures for each mouse nest and individual, marking the first field study to decode wild animal communication in such detail. Mathevon noted that machine learning was essential due to the sheer volume of vocalizations, which humans cannot process manually. Beyond static identity cues, the researchers aim to decode dynamic information, such as stress levels, embedded in the calls. Their work is one of four finalists for the Dolittle Prize, which offers $100,000 for advances in animal communication research. If a team achieves two-way communication with a species, they could receive either $10 million in investment or $500,000 in cash. Mathevon’s broader research spans birds, dolphins, monkeys, and even human babies, challenging the notion that complex communication is unique to humans. His book *The Voices of Nature* explores how animals convey meaning through sound. The African striped mouse study builds on this work, showing that even imperceptible high-frequency calls carry sophisticated social signals. The findings suggest that bioacousticians may soon achieve interspecies dialogue, akin to the fictional Dr. Dolittle. However, experts remain divided on whether such communication would benefit animals or disrupt their natural behaviors. The research highlights the potential of AI to unlock new understandings of wildlife, though ethical implications and practical applications remain open questions.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.