Korea's KICT Unveils AI System That Predicts Toxic Gas Risks in Underground Worksites
The research institute's platform combines IoT sensors with predictive algorithms, estimating gas concentrations in sensor-free zones and pinpointing leak sources from just 30 seconds of data.
The Korea Institute of Civil Engineering and Building Technology (KICT) has announced an underground environmental risk management system that combines IoT sensors with AI prediction to monitor hazardous gases, temperature, humidity and airflow in below-ground workspaces in real time. The system targets hydrogen sulfide, carbon monoxide, volatile organic compounds and fine dust.
Its headline capability is estimating conditions where no sensors are installed: in a demonstration at a large conduit tunnel, the system estimated gas distribution across 14 zones using only two sensors, and identified the source of a leak from 30 seconds of data. The platform was developed through South Korea's science ministry digital twin program for 2025–2026. The announcement marks a shift from reactive gas monitoring toward predictive safety management — an approach project owners, including those in Indonesia's mining and tunnel sectors, are expected to demand as worksite safety standards rise.