Korea is turning to artificial intelligence and robotics to protect its centuries-old cultural heritage, deploying autonomous patrol robots at wooden palaces and sprawling royal mausoleums exposed to physical threats and environmental hazards.

The Royal Palaces and Tombs Center, an arm of the Royal Heritage Administration, announced Friday that it is launching field tests of autonomous AI patrol robots across major historic sites. Building on a brief two-month trial at Changdeok Palace earlier this year, the expanded initiative will run at Deoksu Palace and the Seooreung Royal Tombs through November. The stakes extend beyond national pride: Changdeok Palace has been a UNESCO World Heritage Site since 1997, and Seooreung is part of the Royal Tombs of the Joseon Dynasty, inscribed on the UNESCO list in 2009, placing the sites under international as well as domestic heritage stewardship.

The four-month pilot program will test how different robotic designs handle the delicate, varied terrain of Korea's royal grounds. Traditional security setups often struggle to cover vast outdoor spaces; the autonomous machines will instead roam predetermined routes, using light detection and ranging (LiDAR) sensors and vision-based artificial intelligence to monitor site conditions and assist human guard forces. The choice of platform matters: autonomous security robots have been trialed at airports, shopping malls and corporate campuses worldwide, but historic sites — with fragile surfaces, dense foliage and uneven ground — remain a far less tested frontier for the technology.

The initiative reflects a growing urgency across Korea to fortify vulnerable historical structures against irreversible damage. Protecting wooden palaces and stone monuments has long posed a complex challenge for heritage managers. In 2008, the nation suffered a tragic cultural disaster when an arsonist set fire to Sungnyemun, Seoul's iconic wooden southern gate, reducing the 600-year-old national treasure to ashes; a painstaking restoration took more than five years to complete before the gate reopened in 2013. More recently, high-profile acts of vandalism — including spray-paint graffiti targeting the historic stone retaining walls of Gyeongbok Palace — demonstrated how quickly sprawling cultural landmarks can be defaced despite conventional camera surveillance.

To suit differing ground topographies, authorities are deploying distinct robotic models. A four-wheeled driving robot will navigate the paved, relatively flat courtyards of Deoksu Palace in central Seoul. A bipedal walking robot will traverse the uneven paths, unpaved trails and steep inclines of Seooreung, a sprawling forest complex housing royal tombs in Goyang, northwest of the capital.

Equipped with optical and thermal imaging cameras, the robots are programmed to detect real-time emergencies — such as fires, unauthorized trespassers, or medical crises in which visitors collapse. Connected directly to a centralized security control center, the units can deliver automated warnings, broadcast safety announcements and immediately alert human response teams.

Preserving timber-framed palaces and expansive royal burial grounds presents unique logistical challenges: narrow passages, dense foliage and delicate historical artifacts limit the installation of fixed surveillance hardware. By deploying adaptive autonomous systems, heritage officials said they aim to evaluate whether walking or wheeled platforms offer better protection for vulnerable cultural assets. The Korean pilot builds on a broader shift in which cultural heritage bodies worldwide have adopted remote sensors, drones and AI-based monitoring to safeguard fragile sites — and it draws on a country that has ranked first globally in industrial robot density for more than a decade, according to the International Federation of Robotics.

If the trial proves successful, officials plan to extend the program through next year, laying the groundwork for a broader, technology-driven safety and conservation framework across Korea's most cherished historical landmarks.

This article was published with the assistance of generative AI and edited by The Korea Times.