A groundbreaking study has revealed that African lions produce two distinct types of roars, not just one as previously believed. This discovery, led by researchers at the University of Exeter, could significantly enhance conservation efforts by improving the accuracy of lion population monitoring. The team identified a previously unknown "intermediary roar" that occurs alongside the well-documented full-throated roar, challenging long-held assumptions about lion vocalizations. Published in the journal Ecology and Evolution, the research marks the first time artificial intelligence has been used to automatically classify lion roars into different categories. The machine learning system achieved a 95.4% accuracy rate, drastically reducing the subjectivity associated with human analysis and enabling more consistent identification of individual lions. Lead author Jonathan Growcott explained that lion roars function as unique acoustic signatures, much like fingerprints, which can be used to estimate population sizes and track individual animals over time. "Until now, identifying these roars relied heavily on expert judgment, which introduced potential bias," he said. "Our AI-driven approach offers a more objective, accurate, and scalable method for monitoring lion populations—critical for protecting a species that is increasingly at risk." Lions are currently listed as vulnerable to extinction by the International Union for Conservation of Nature. Estimates suggest only 20,000 to 25,000 wild lions remain in Africa, a decline of about 50% over the past 25 years. The new findings suggest that a more nuanced understanding of lion vocal behavior is essential for effective conservation. The study’s success highlights the growing importance of bioacoustics in ecological research. Similar advances have been observed in studies of other large carnivores, such as spotted hyenas, underscoring the potential of sound-based monitoring across species. By using AI to analyze passive acoustic data, researchers can now process vast amounts of audio recordings more efficiently than traditional methods like spoor surveys or camera trapping. This not only increases the speed and scale of data collection but also reduces costs and human error. Growcott emphasized the need for a shift in wildlife monitoring practices: "We believe there must be a paradigm shift toward large-scale adoption of passive acoustic techniques. As bioacoustic technology continues to improve, it will become indispensable for conserving lions and other threatened species." The research was a collaborative effort between the University of Exeter, the Wildlife Conservation Unit at the University of Oxford, Lion Landscapes, the Frankfurt Zoological Society, TAWIRI (Tanzania Wildlife Institute for Research), and TANAPA (Tanzania National Parks Authority). Computer scientists from both universities played a key role in developing the AI models. Funding was provided by the Lion Recovery Fund, WWF Germany, the Darwin Initiative, and the UKRI AI Centre for Doctoral Training in Environmental Intelligence.