|  Programme

Since 2015, the Ministry of the Interior and Safety has annually published the Regional Safety Index. The Regional Safety Index uses various safety-related statistics to assess the safety capacity of local governments across six safety categories: traffic accidents, fires, crime, daily-life safety, suicide, and infectious diseases. The results are presented on a scale from Grade 1 to Grade 5. 

The indicators used to calculate the Regional Safety Index consist of hazard indicators, vulnerability indicators, mitigation indicators, and awareness indicators. Hazard indicators refer to the number of deaths in each category. Vulnerability indicators refer to human and physical factors that contribute to the occurrence of hazards. Mitigation indicators refer to local governments’ efforts to prevent and respond to such incidents, while awareness indicators refer to residents’ efforts to prevent and respond to them. 

After the Regional Safety Index is announced, local governments establish and implement plans to improve their safety index scores in each category. To support these efforts, the Ministry of the Interior and Safety has formed a consulting group to assist with regional safety assessments. It also publishes a collection of customized projects by category to help ensure that appropriate initiatives are implemented. 

Beginning this year, the Ministry of the Interior and Safety plans to introduce “industrial accidents” as a new category, expanding the Regional Safety Index from six to seven categories. The ministry is currently in consultation with relevant organizations regarding this change. 

Artificial intelligence (AI) is playing an increasingly important role in disaster risk management, particularly in disaster prevention and response. This topic explores how AI technologies enable real-time risk detection, predictive analysis, and decision support by integrating multi-source data such as social media, sensor networks, and geospatial information in Taiwan. 

Key applications in Taiwan include,
(1) risk detection
(2) risk assessment
(3) automated disaster reporting
(4) the use of GeoAI for spatial analysis.
(5) chatbots with large language models

AI agents have significantly supported disaster risk management, enabling a transition toward human–AI collaboration. However, building trustworthy AI remains a key challenge, including issues such as data bias, privacy protection and security, system transparency, and explainability.

-