South Korea to Pilot AI System for Managing Intersection Congestion

The Korean National Police Agency announced on the 24th that it will launch a pilot project for an artificial intelligence system in December, deploying AI-based traffic cameras in Seoul to address intersection congestion.

According to South Korean media reports, this pilot project, named the "Unattended Intersection Tailgating Enforcement System," will be conducted from December to February next year at an intersection in Gangnam District, southern Seoul. The system will issue warnings to vehicles that enter the intersection when the traffic light is green but remain stranded after the light turns red, thereby obstructing traffic, excluding vehicles forced to stop due to unavoidable emergencies such as traffic accidents or vehicle breakdowns.

The new system employs AI-based video analysis technology developed through a research project by the Korean National Police Agency. The police stated that the relevant technology improves law enforcement accuracy, allowing a single device to simultaneously handle violations like running red lights, speeding, and tailgating, thereby reducing maintenance requirements and improving efficiency.

South Korean police plan to install similar traffic cameras at 10 chronically congested intersections by 2026 and expand the system to 883 congestion-prone intersections nationwide by 2027.

According to South Korea's current "Road Traffic Act," when drivers approach a signal-controlled intersection, they must not enter if there is congestion ahead that could cause their vehicle to stop within the intersection and block other vehicles. Violators will be subject to fines.

"Entering an intersection just because the light is green, or acting on the selfish thought of 'me first,' threatens public safety," an officer from the Korean National Police Agency said. Police advise drivers to assess the congestion situation ahead when the light turns green at an intersection before deciding whether to proceed.