Wolf Acoustics: AI Deciphers Pack Communication
Computational linguist Jeffrey Reed uses machine learning to filter environmental noise and isolate specific wolf howls, aiming to map meaning to sound in Yellowstone's wild canids.
By Yasmin Nasser ·
Jeffrey Reed is using artificial intelligence to strip away background noise from recordings of wild wolves, a process that allows researchers to isolate individual vocalizations for analysis. Reed, a computational linguist with a Ph.D. in ancient languages, returned to his hometown on the outskirts of Yellowstone National Park to apply three decades of Silicon Valley experience to bioacoustics. He previously spent time building machines designed to replicate human language before pivoting his focus to animal communication.
Reed co-founded The Cry Wolf Project, which is described as the world's largest bioacoustics study of wild wolves. The initiative relies on technology developed through Grizzly Systems, another company Reed co-founded that produces hardware for monitoring wildlife. The primary challenge in this field is separating the target signal from the ambient noise of the environment. AI models are now being used to perform this separation, enabling the isolation of specific howls that would otherwise be indistinguishable in raw field recordings.
The project aims to decode the meaning behind each howl, treating wolf vocalizations as a structured language system. Reed’s approach combines his background in ancient languages with modern machine learning techniques. While the technology is currently in a phase of isolating and recording sounds, the ultimate goal is interpretation. This work represents a shift from passive observation to active decoding of complex animal behavior.
Reed has a book on wolf communication forthcoming from Little, Brown and Company. The research is part of a wider conversation on whether curiosity and technology can aid in preserving biodiversity. A related TED Talk explores the possibility of learning to talk to sperm whales, while another examines the imitation of nature’s voices. These projects suggest a growing intersection between computational linguistics and conservation biology.
The segment discussing Reed’s work was produced by Phoebe Lett and edited by Sanaz Meshkinpour for the TED Radio Hour. The episode is part of a series titled Can curiosity save the wild? Reed’s work stands at the intersection of language processing and field ecology, providing new tools for understanding how wolves communicate in their natural habitat.
