Wildlife conservation is undergoing a profound transformation thanks to artificial intelligence. While traditional methods of monitoring lion populations relied on GPS collars or camera traps, a major scientific breakthrough now shows that a lion can be identified solely by the sound of its roar. This discovery marks a turning point in the field of AI-assisted bioacoustics, where animal vocalizations are becoming a new and essential source of data for conservation1.
An unexpected discovery among more than three thousand recordings
The study, led by Jonathan Growcott, a doctoral student at the University of Exeter, utilized fifty microphones deployed in Nyerere National Park in Tanzania, as well as five acoustic collars fitted to lions in the Bubye Valley Reserve in Zimbabwe. The goal was to refine population counting methods using vocalizations, but analysis of the recordings revealed an unexpected result.
In addition to the well-documented powerful roars and internal communication calls, researchers have identified a third category: an intermediate roar that is deeper and more uniform. Previously inaudible to human observers, it was successfully isolated using a deep learning model that achieved a classification accuracy of over ninety-five percent.
AI reveals a unique vocal signature specific to each lion
Each powerful roar contains a unique acoustic signature that allows a lion to be identified individually. Automated analysis has made it possible to:
- identify more than 3,000 records,
- achieve a classification accuracy of over 95%,
- identify gender, age, and certain physiological markers from the audio signal,
- identify a recurring vocal pattern in which the intermediate roar consistently follows the powerful roar.
These findings show that lions have a more complex vocal repertoire than previously documented2.
Practical applications for conservation and anti-poaching efforts
The implications for conservation are significant. This approach makes it possible to:
- non-intrusive monitoring of lions,
- more accurate acoustic mapping of regions,
- remote detection of lions in hard-to-reach areas,
- a significant improvement in population estimates,
- increased support for anti-poaching programs, including nighttime surveillance and the detection of individuals near protected areas.
With the global lion population estimated at around 23,000 individuals in the wild, AI is becoming a crucial tool for supporting conservation efforts3.
Ethical Issues and Limitations of a Rapidly Evolving Technology
Like any method that relies on the collection of sensitive data, AI-assisted bioacoustics requires strict oversight. Reliability depends on the quality of the recordings; analyses must be biologically contextualized; and the technology must not be used outside its scientific or conservation framework. Researchers also emphasize that animal vocalizations must be studied within a comprehensive ecological framework that integrates genetics, the environment, and social behaviors.
Conclusion: Toward an Ecology Enhanced by Artificial Intelligence
This breakthrough opens up new possibilities for the conservation of big cats. AI now makes it possible to detect acoustic signatures inaudible to the human ear and to transform monitoring methods into tools that are more accurate, less intrusive, and better suited to the study of endangered species. As ecology relies on more sophisticated technologies, artificial intelligence is emerging as an essential ally in addressing global biodiversity challenges.
Learn more
To explore another major application of artificial intelligence in conservation, see: Ocean AI: How Artificial Intelligence Is Helping to Protect Whales in Polynesia
References
1. University of Exeter. (2025). Acoustic Identification of African Lions Using Deep Learning Models.
https://www.exeter.ac.uk/research
2. Ecology and Evolution. (2025). Bioacoustic Patterns in Panthera leo and AI-Assisted Vocal Recognition.
https://www.onlinelibrary.wiley.com/journal/20457758
3. International Union for Conservation of Nature. (2024). African Lion Population Status and Conservation Challenges.
https://www.iucnredlist.org/species/15951/115130419