Russia Develops an AI‑Powered Audio System to Monitor Wildlife Populations
Russian researchers have created an AI system capable of recognizing wolf howls with high accuracy—opening new possibilities for non‑invasive wildlife monitoring and ecosystem management.
AI Learns to Hear the Wild
Scientists from the Severtsov Institute of Ecology and Evolution, the Vavilov Institute for the History of Science and Technology, and Sber have developed a neural‑network method for automatically detecting wolf howls.
Special acoustic traps capture environmental sounds, while the Audio Spectrogram Transformer (AST) analyzes them. The intelligent detector identifies all recorded audio events and isolates wolf howls among them. The system recognizes animal vocalizations with an accuracy of up to 89.6 percent.
Smart Tools for Conservation
The smart acoustic trap is designed for ecological monitoring and wildlife protection. It allows researchers to estimate wolf populations, a task that is traditionally difficult due to the animals’ secretive behavior and the labor‑intensive nature of classical field methods.
Wolves play a vital role in ecosystems by helping regulate populations of hoofed animals. Audio‑based monitoring can also be extended to other species.
Across Russia, digital tools—ranging from AI systems to drone‑based monitoring—are becoming essential assistants to conservationists, enabling more effective tracking of wildlife in remote and protected regions.
