From caws to code: AI helps decrypt animal communication
Researchers in northern Spain are using AI models from the Earth Species Project to decode the complex vocalizations of cooperative-breeding carrion crows, identifying over 127,000 calls and distinguishing between adults and juveniles. The collaboration aims to synchronize audio data with behavioral observations to understand how these birds coordinate their activities.
Scientists Vittorio Baglione and Daniela Canestrari at the University of León in northern Spain have spent decades studying how carrion crows communicate, particularly in cooperative-breeding groups where entire families raise chicks. The researchers deployed audio recorders and biologgers to capture crow calls but faced overwhelming data volumes, with each microphone lasting only six to seven days. In 2024, they partnered with the U.S.-based Earth Species Project (ESP), a nonprofit developing AI tools to analyze animal communication. ESP’s models helped categorize over 127,000 crow vocalizations, differentiating between adults and juveniles and synchronizing data from multiple loggers. Initial findings suggest most calls are soft, low-amplitude murmurs, indicating close-range communication. The team is now integrating audio data with video footage and accelerometer readings to map behaviors and call functions. ESP’s generalizable AI model, adaptable across species, also supports global research, including studies on orcas and underwater noise impacts. Senior AI researcher David Robinson notes that understanding animal communication could deepen insights into biodiversity and species complexity. While data analysis continues, the project aims to create a semantic map linking audio, video, and movement data to decode crow communication patterns. The collaboration highlights AI’s role in accelerating conservation research and bridging humanity’s disconnect from nature.
This content was automatically generated and/or translated by AI. It may contain inaccuracies. Please refer to the original sources for verification.