The Pulse

South Korea Plans $692 Million for AI Disaster Response

South Korea plans to invest 940 billion won, about $692 million, in 50 AI-related disaster and safety projects through 2030. Three local governments will test AI services that analyze storm risks and recommend evacuations and road closures.

AI.info Team · Written with AI · fact-checked · 2026-09-28

South Korea wants to move disaster management from responding to emergencies toward anticipating them, committing about 940 billion won—roughly $692 million—to 50 AI-related projects through 2030. The shift puts predictive tools into a system the government describes as centered on response, but the first test will be narrower: pilot services for heavy rain and typhoons in three local governments.

The Ministry of the Interior and Safety announced the plan on September 27. It says AI will combine information now held across agencies to assess hazards and help officials decide when to evacuate people or restrict access to dangerous areas. The ministry has not presented the plan as an automated emergency-response system; its pilots are designed to recommend actions and support officials’ decisions.

Three local governments will test the storm service

Busan, Yeongcheon in North Gyeongsang Province and Inje County in Gangwon Province were selected through a public competition this year. Their pilot services will analyze real-time weather, risks such as underground spaces and low-lying areas, and records of previous disaster damage. The system is intended to recommend suitable evacuation and access-control measures during typhoons and heavy rain.

The ministry says the project’s launch meeting took place on September 8. It plans to work with the three local governments to assess the decision-support service by early 2027. The announcement does not provide performance results or say how the service will be evaluated beyond that planned effectiveness review.

AI depends on joining records held by separate agencies

Before those recommendations can be useful, the government plans to convert existing disaster records into formats AI systems can process. The ministry and the National Information Society Agency will structure material such as daily safety-management reports and emergency notification messages, which are currently unstructured documents. The plan also calls for combining satellite information with data about areas vulnerable to natural hazards and landslides.

For a heavy-rain event, the ministry’s example brings together flood extent and damage records, weather information, river levels and 119 emergency calls. Those records come from different agencies, including the Interior Ministry, the Korea Meteorological Administration, the Ministry of Climate, Energy and Environment, and the National Fire Agency. Linking them by disaster event is meant to give officials a more complete view of changing conditions.

The plan still puts officials between forecasts and action

The government’s stated aim is to replace a response-centered approach with AI-supported preparation. That makes the quality and availability of shared data as important as the recommendations themselves: the ministry says it will also set out procedures for identifying demand for disaster-safety data and providing it to public institutions and private companies. It also plans to encourage AI-friendly designs when public websites are newly built or redesigned.

Interior and Safety Minister Yoon Ho-jung framed data preparation as a practical requirement for field staff. “We will quickly convert and integrate the disaster-management data accumulated over the years into forms that AI can understand,” he said.

The near-term measure is whether the three pilot services can help local officials make better-supported decisions during severe rain and typhoons. The ministry’s planned review by early 2027 will be the first stated checkpoint for that test.