On February 27, 2026, global big tech company Google received conditional approval from the South Korean government to export Korea’s 1:5000 map data. It came a full 19 years after its first request. Maps we now hold in one hand have never once been unimportant from antiquity to today. Now maps are evolving beyond scale into a new dimension.
“The turn-by-turn navigation feature on Google Maps is not available in Korea alone.”
In August 2025, Chris Turner, Google’s Vice President for Government Affairs and Public Policy on Knowledge and Information, wrote this on the Google Korea blog. At the time, Google had just filed its third request with the Korean government to export 1:5000 map data.
Back in 2007, when Google was expanding its Google Maps service worldwide, it first requested access to Korea’s 1:5000 National Base Map data. It was not simply asking to view or use the data locally; it was demanding permission to move the data to its overseas servers for storage and processing.
As a result, the core of the controversy became the concept of “export” itself, and for the next 19 years it turned into a tug-of-war spanning policy, industry, and diplomacy.
“The reason export was denied in 2007 was Korean law.” So said Im Si-young, Associate Research Fellow at the Spatial Information Policy Research Center of the Korea Research Institute for Human Settlements. Korea’s map data are location information assets created with national budget funds, and at the time they were legally prohibited from being taken out of the country.
“Legally, there was simply no way to provide them.”
Things began to change around 2016. By then, with services such as Google Maps and Google Earth providing two-dimensional (2D) maps and aerial/satellite imagery worldwide, map data were starting to be recognized as core infrastructure for industry and a wide range of services. So when Google again requested data export, the Korean government shifted from a stance of “blanket prohibition” to one of allowing export if three specific conditions were met.
The first condition presented at the time was national security. “The first condition was a request to process satellite and aerial imagery so that key national facilities, such as military bases, could not be identified,” Im explained.
In particular, it was seen as problematic that the locations of military bases could be clearly seen in the Google Earth service based on satellite imagery. Google refused, arguing that such processing would distort the truth.
The second condition was to restrict the exposure of precise location coordinates. Im explained, “Map data are no longer just images — once coordinates and attribute values are combined, they become ‘spatial data,’ and three-dimensional (3D) coordinates can then be used as reference points for precision missile strikes.”
The final condition was to process the data within Korea. “If Korea provides map data and the Dokdo islets end up labeled as ‘Takeshima,’ from Korea’s perspective that is a kind of ‘incident.’” Google did not accept such demands for roughly 10 years.
In addition, there was strong public sentiment emphasizing sovereignty and protection of domestic industry. Map data can be processed and reused across various industries and technologies to generate new value. This fueled concerns that moving data to foreign platforms could lead to a loss of control over how those data are used.
There were also worries that the competitive landscape between global giants and Korean companies such as Naver and Kakao could change dramatically. The points of contention shifted over time: in the early years, security was paramount; in later years, data sovereignty and protection of domestic industry came to the fore. So what exactly is this 1:5000 map data that kept society in an uproar for so long?
After Google’s second request to export map data in 2016, the Korean government asked the company to prevent the identification of key national facilities and the disclosure of precise coordinates. Google did not accept these terms at the time. Even today, the exact location of the Blue House can still be seen in satellite imagery on Google Maps. Getty Images Bank
• Everything on the Korean Peninsula, in 1:5000 map data
A scale of 1:5000 means that 1 cm on the map represents 50 m in the real world. With a scale like 1:25000, the larger denominator means that the same 1 cm represents a greater distance — 250 m. Such maps can cover a wider area, but at the cost of less detail for each feature.
Lee Sang-il, professor in the Department of Geography Education at Seoul National University, explained, “At a scale of 1:5000, the map not only shows locations, it also captures the structure of space.” In other words, beyond simply recognizing whether a road or building exists, you can see details such as the shape and arrangement of buildings and the structure of roads, revealing how the space is organized.
In Korea, the 1:5000 map serves as the National Base Map. A National Base Map is a standard reference map created and managed by the state, forming the starting point for all maps and spatial information. In Korea, it is produced by the National Geographic Information Institute.
Kim In-hyun, CEO of Korea Spatial Information & Communication, who first informed the media about Google’s data export request in 2007, explained, “Everything from mainstream consumer map services like Naver Map or Kakao Map to national policies in urban planning, disaster response, and infrastructure management relies on the National Base Map.” The National Base Map is the fundamental ingredient for all services that utilize maps.
Worldwide, only about five countries — Korea, Japan, Spain, France, and Germany — possess 1:5000 scale maps of their entire territory. In countries that do not produce 1:5000 maps, Google purchases satellite imagery to provide its services.
The National Base Map produced by the National Geographic Information Institute is not a single, flat image. It is a multi-layered dataset that represents a single space by splitting it into multiple information layers.
“The National Base Map is vector data with up to 107 layers,” Kim said. “We separate elements such as road networks, buildings, terrain, rivers, administrative boundaries, and facilities into independent layers. When you stack them on a single coordinate system, you get what we commonly think of as a ‘map.’”
Thanks to this multilayer structure, the National Base Map becomes computable data. For example, you can find the closest road from a specific coordinate, calculate the distance between a building and a road, or select areas that satisfy particular conditions. In effect, the National Base Map contains virtually all spatial data about the Korean Peninsula.
On Google Maps in Korea, turn-by-turn directions for car, walking, and bicycle navigation are not provided (red box), except for public transit. That is because providing such map-based services requires access to National Base Map data. Screenshots: Google Maps, Kakao Map
When asked, “How are map data usually used?” Im answered, “Unless you make a deliberate effort not to use them, they are used in virtually every industry and field.” He added that the terms “map” or “map data” can actually make people underestimate the breadth of their applications, emphasizing that they should be understood not as simple pictures but as “spatial information,” combining locations, objects, and relationships.
In February 2026, the Korean government approved the conditional export of 1:5000 map data. This followed Google’s third request in 2025, after which the government asked the company to supplement technical details regarding image security processing, coordinate display restrictions, server location, and post-management. The approval came after reviewing these measures.
On February 27, 2026, the Ministry of Land, Infrastructure and Transport announced in a press release that it had “decided to approve the export on the premise of strict compliance with security conditions.” According to the release, even after map services go live, Google must manage procedures to implement government-requested corrections and must station a dedicated map liaison in Korea to maintain an always-on communication channel with the government.
However, because Google still does not maintain servers in Korea, the issue of corporate tax avoidance is expected to continue. As of 2024, Google does not have a permanent establishment in Korea and pays only about 4.3% of the corporate taxes paid by Naver and Kakao.
“Google Maps is not going to change dramatically just because of the export,” Im noted, pointing out that Google is already partially using 1:5000 map data provided by Korea’s navigation service “T map.”
“Turn-by-turn directions for cars, walking, and bicycles will be added soon,” Professor Lee predicted. He explained that providing map-based services in Korea requires building on the National Base Map and going through the government’s approval process, which has constrained Google’s services.
In this way, the 19-year standoff with Google has come to a close. Yet experts say that now is precisely the time to “pay attention to the value of map data.” After all, maps that humans have been creating for thousands of years are now transforming from 2D to 3D.
Nineteen years after first requesting the export of high-precision 1:5000 map data, Google finally received approval in 2026, on the condition that it comply with the Korean government’s requirements for security and data management. Getty Images Bank
• From 2D to 3D: the rise of the digital twin city
“This is a 3D re-creation of the Saudi Arabian cities of Mecca, Jeddah, and Medina.”
On April 3, at Naver Labs’ headquarters in Pangyo, Vision Group leader Lee Dong-hwan pulled up a screen. An exotic urban panorama appeared. Jeddah is Saudi Arabia’s second-largest city, and Mecca and Medina are historic Islamic holy sites. All three are massive cities built in the desert. What was striking was that the city on the screen was not just a bird’s-eye 3D map, but “data” that replicated the city itself.
In the past, 3D maps were little more than an “expansion of representation,” showing the height and shape of buildings in three dimensions. Today’s 3D maps have evolved into “digital twins” — three-dimensional digital counterparts that replicate the structure, state, and changes of the real world in real time and even support interaction.
On Lee’s screen, a flood simulation of the city started playing.
“One of the main reasons Saudi Arabia first reached out to Naver Labs was for flood simulation. It might sound odd to talk about flooding in the hot Middle East, but because of insufficient drainage infrastructure, heavy rains can turn streets into rivers.”
Together with the Korea Water Resources Corporation, Naver Labs built a system that uses 3D maps to simulate heavy downpours: which paths the water will follow, where it will pool, and which buildings will be inundated.
A digital twin view of Seoul created by Naver Labs in 2020. Digital twins accurately reproduce reality and help improve the precision and efficiency of policy decisions. Courtesy of Naver Labs
3D maps include elements such as building heights and shapes, terrain slopes, and road structures. On a flat map, it is difficult to predict exactly how water will flow or where it will accumulate, but a 3D map makes it possible to calculate water movement based on the actual terrain and built structures.
And it’s not just floods. 3D maps can be used to predict and plan responses to other disasters, as well as traffic and environmental changes. In effect, 3D maps create a “digital twin city” — a virtual city model that replicates the real city as 3D data.
3D maps are generated in the form of “point cloud data,” which densely sample space point by point by combining aerial photography, LiDAR sensor data, and other sources. As a result, they become data that reveal how a space is configured and how it operates.
In particular, technologies already emerging today and expected to expand further in the future — autonomous vehicles, physical AI, and augmented reality (AR) — are all projected to develop on the foundation of 3D maps. 3D map data produced as point clouds are used for robots to compute travel routes, understand road structures, and enable AR devices to recognize their location.
• A 1:1 scale? Future maps become reality itself
“The expansion of maps into 3D is a tremendous leap forward,” Professor Lee said. In particular, whereas traditional maps were limited to outdoor information, 3D maps are aiming not only to include indoor spaces but also to link outdoors and indoors seamlessly.
Lee explained that “our perspective on the world through maps will change.” Yet a transformation this significant does not promise only positive outcomes.
“Once built, a 3D map becomes data that almost perfectly reflects the state of reality.”
CEO Kim voiced concern about 3D maps. We are moving toward data that combine 3D building structures and spatial forms with the movements of people and objects.
Lee of Naver Labs noted, “It’s important to comply with domestic laws on point cloud data while using it as national infrastructure to foster industry,” adding, “From a security standpoint as well, we need to build 3D maps using domestic companies’ technologies.” Map data could again become the focus of debate over sovereignty and industrial protection.
“The biggest concern is the misuse of information.” Im pointed out that 3D maps can be abused in two ways: for security threats and privacy violations. Because 3D maps incorporate building heights, structures, and positions combined with precise coordinates, they offer significant potential for military use.
Furthermore, 3D maps that seamlessly connect building exteriors and interiors can trigger serious privacy issues. A fully realized 3D map model could do more than reveal interior layouts through open windows; it could infer the indoor temperature and how energy circulates within the building.
“For now, however, 3D mapping is focusing on public spaces, and for high-security facilities, not all spatial information is disclosed,” Im explained.
On the other hand, 3D maps may not be built solely for humans. “Above all, 3D maps will converge with artificial intelligence (AI),” Professor Lee predicted, expecting 3D maps to play a major role in AI’s path toward artificial general intelligence (AGI).
“Initially, we trained AI models solely on language or solely on images. Eventually, AI will be able to recognize everything humans see and sense. There is nothing better than a digital twin for training such models. We need to imagine the impact when 3D maps evolve beyond powering specific industries and instead become data used by ‘everyone.’”
The future of maps calls to mind “On Exactitude in Science” (“Del rigor en la ciencia”), a short story written in 1946 by Argentine writer Jorge Luis Borges. In the story, cartographers, striving for ever greater accuracy, ultimately create a map at a 1:1 scale. That map, the same size as reality itself, proves useless and is abandoned in the desert.
Could Borges have imagined that the 1:1 scale map he conceived as a satirical fable would one day be realized under the names of spatial information and digital twins? We are still living in the age of maps.
*Related article (in Korean)
Science Donga, May issue, [Current Affairs Feature] From 1:5000 to 1:1, from 2D to 3D: The Future of Maps and Maps of the Future
Copyright ⓒ DongA Science. All rights reserved.