Ground truth: When the Earth moves, AI can spot it
Airfind news item
By Chris Baraniuk
Published on March 24, 2026.
The University of Melbourne's artificial intelligence (AI) system has developed a map created by machine learning to identify potentially high-risk areas in which the ground is shifting. The map was created using an AI system that identified an unstable area beneath the village of Kimtang in central Nepal, which is on the edge of a potentially landslide. This map highlighted an area on the hillside particularly particularly unstable, marking it as one of the most stable spots in the unstable area. The system was developed using advanced machine learning techniques, which can identify slopes at high risk of slipping. The data was obtained via a European satellite called Sentinel-1, which used radar to map terrain in fine detail. Despite this, no significant incidents have occurred since the AI was able to identify these areas. Researchers are working with the local authorities to monitor the situation and develop ways of monitoring and evacuation routes. However, this information and analysis could not be made without their deep knowledge of the process.
Read Original Article