Mongabay · Nepal
Local knowledge guides an AI tiger-warning pilot in Nepal
Recap published
Indigenous communities in Bardiya were already listening to the forest before researchers arrived with sensors: alarm calls from deer, langurs and rhesus macaques can signal that a tiger is near. Researchers trained a low-cost AI on chital deer alarm calls to map areas people should avoid. Forest guard Thenghchu Tharu compared its alerts with signs on the ground: "It matched exactly." The study notes false detections, and the pilot ended after its trial for lack of funding, staff and local institutional support.
Why it matters
Knowledge that residents already used was built into an early-warning pilot, and a forest guard who holds that knowledge compared the results with what he saw.
Origins, evidence and credits
Original headline: In Nepal, AI builds on local knowledge to help prevent tiger attacks
Original reporting: Mongabay ·
- Origins and holders
- Indigenous communities in Bardiya used forest alarm calls before the sensor pilot. The recap credits people in the roles reported by Mongabay.
- What was checked
- The article reports that forest guard Thenghchu Tharu compared system alerts with signs on the ground.
- Limits reported
- The study notes false detections. The pilot ended after its trial without the funding, staffing and local institutional support needed to continue.
People and institutions credited
Bhadai Tharu (chair, Gauri Community Forest, interviewed); Thenghchu Tharu (forest guard, Dalla Community Forest, who helped monitor the pilot); Ramjan Chaudhary (field biologist, Bardia Nature Conservation Club; study co-author); Casey Anderson (study co-author, who led the deployment). Study: Kershenbaum et al. (2026), *Remote Sensing in Ecology and Conservation* 12(4), 661–672, doi:10.1002/rse2.70061.