National Library of Korea Releases Text Data for "Homegrown AI" Training...Transforms into a Data Hub

Copyright-cleared holdings to support the building of an independent AI model
Expanding the realm of knowledge resources, including acquisition of 88,000 pages of records from the U.S. NARA

The National Library of Korea is set to transform from a simple book repository into a data hub that supplies core resources for the national artificial intelligence (AI) industry. It will process holdings whose copyrights have expired or been cleared into high-quality text data and open them to the private sector and the government.

The National Library announced on the 4th its "2026 Key Work Implementation Plan," which includes these initiatives. The core of the plan is to support the government’s ongoing project to build an independent AI foundation model. Refined Korean-language data is essential to develop a "homegrown AI" that accurately understands the context and nuances of Korean. To this end, the library will cooperate with the Ministry of Science and ICT to convert its vast holdings into AI training data and provide them. Through the "Shared Library" section on its website, these resources will be fully opened to support the creation of K-content and technology development.

The library will also expand the scope of knowledge resources that serve as the source of this data. This year, it will newly collect more than 330,000 volumes of printed materials and about 400,000 items of digital materials, including e-books, to increase the total volume of national knowledge. In particular, it will discover independent publications that are easily excluded from commercial distribution channels through crowdfunding platforms, and include non-standard publications such as conference materials from international events in its collection targets.

An official from the National Library said, "For overseas Korean studies materials, we will expand the scope of our surveys from the existing focus on the United States and Japan to include the United Kingdom and Germany, and we plan to strengthen research materials on modern and contemporary Korean history by acquiring more than 88,000 pages of Korea-related records held by the U.S. National Archives and Records Administration (NARA)."

The library will also strengthen its educational role so that the public can utilize the resources and technologies it has secured. It will expand its AI literacy education courses from the current 25 sessions to 32 sessions, and provide young people with programs that combine new technologies and reading, such as augmented reality (AR) musicals and an AI digital ethics experience center.

The library will accelerate exchanges with the international community as well. In particular, during the "2026 World Library and Information Congress" to be held in Busan this August, it will host the "National Librarians Meeting," which will bring together the heads of national libraries from more than 100 countries around the world. International discussions will be held on the role of libraries in the AI era and on data sovereignty.

Kim Heeseop, Director General of the National Library of Korea, said, "This year will be the year when the library’s knowledge resources combine with AI technology to create tangible value," adding, "As the nation’s representative library, we will focus on securing data and innovating our services."

This content was produced with the assistance of AI translation services.