# Ground truth: When the Earth moves, AI can spot it

## Introduction
• In central Nepal, the village of Kimtang shows signs of instability due to ground movement, raising concerns about potential landslides.
• The use of AI technology to analyze satellite imagery can predict landslide risks, which could save lives and inform evacuation plans.

## Key Developments
• Mathematician Antoinette Tordesillas from the University of Melbourne utilizes AI to highlight unstable areas beneath Kimtang, marking them in red on radar maps.
• Landslides claim thousands of lives and cause billions in damages each year, particularly exacerbated by climate change and human activities like construction.
• Statistics show that landslides in the US result in 25-50 fatalities annually and vast financial losses.

## How It Works / Details / Explanation
• Satellite images captured using radar can reveal ground movements that are not visible to the naked eye, identifying cracks and subtle separations in soil.
• AI, specifically machine learning, processes and analyzes large data sets generated from satellite imagery, something beyond human capability.
• The analysis from Sentinel-1 satellite enabled the research team to inform villagers about the risk and establish monitoring protocols.

## Context / Background
• Nepal’s geographical location in the Himalayas makes it particularly susceptible to landslides; a significant incident in October 2025 claimed around 60 lives.
• The advancements in AI are critical for developing predictive measures for geological hazards, with scientists utilizing historic data to identify high-risk slopes.

## Implications / Impact
• AI’s predictions allow for timely warnings, potentially reducing casualties and economic costs associated with landslides.
• Similar applications of AI are being used in the UK, where researchers analyze 300,000 slopes, discovering 3,000 that are actively moving and presenting future risks.
• The data shared with local authorities aids emergency responses and infrastructure management, enhancing public safety.

## Challenges / Risks / Criticism
• The reliance on AI systems presents the challenge of false positives, which can lead to unnecessary panic or misallocation of resources as noted during avalanche detection trials in the Alps.
• The need for continuous real-time data remains a barrier, as accessing the latest satellite information can be prohibitively expensive.

## Future Outlook
• The integration of AI technology in geological monitoring is set to expand, potentially incorporating real-time data in collaboration with various global research institutions.
• Ongoing projects may improve algorithms and reduce error rates, increasing the effectiveness and reliability of landslide predictions.

## Conclusion
• AI technology represents a groundbreaking advancement for predicting landslides, showcasing the potential of data science in enhancing public safety and mitigating disaster impacts.