Hong Kong develops AI-driven landslip warning system with more than 90% accuracy
New system, to be implemented next year, boosts landslide risk management by combining AI, big data, and real-time rainfall monitoring
Hong Kong authorities have developed an AI-based landslip warning system, increasing the accuracy of predicting the number of landslides to more than 90 per cent and improving risk assessment and warning issuance during adverse weather.
The Geotechnical Engineering Office of the Civil Engineering and Development Department said on Tuesday that it aimed to fully implement the new system next year. It will allow real-time assessment of landslide risks to optimise the city’s warning procedures.
Raymond Cheung Wai-man, head of the office, said the new artificial intelligence (AI) model, the fifth generation of the government’s landslip warning system, offered enhanced prediction accuracy compared to earlier versions.
It can increase accuracy in predicting the number of landslides to more than 90 per cent, compared with nearly 70 per cent by the current system.
The new system also features slope-specific prediction models tailored to Hong Kong’s geological conditions. It integrates more factors contributing to landslides, such as rainfall characteristics and man-made slopes, into its analysis.
“Early warning is a crucial component in both responding to extreme weather events and even in Hong Kong’s slope safety system,” Cheung said.
“By leveraging new technologies, we hope to better understand Hong Kong’s overall landslide risks.”