The government has decided to move up the release of its ‘high-value public data TOP 100’—datasets with strong potential for artificial intelligence (AI) development—from the originally planned 2028 to 2027, a full year ahead of schedule. It will also open 37 categories of public works free of copyright and personal-information concerns so that all AI companies can freely use them for training.
 [코리안투데이] Prime Minister Han Seong-sook chairing the Data-related Ministers Meeting © 장형임 기자 |
On August 27, the government held a Data-related Ministers Meeting at the Government Complex in Seoul, chaired by Prime Minister Han Seong-sook, to discuss ways to expand the release of public data. It was the second session of the Data-related Ministers Meeting, established in May as a whole-of-government body overseeing data policy. Ministers and vice ministers from the Ministry of Science and ICT, the Ministry of the Interior and Safety, the Ministry of Culture, Sports and Tourism, the Ministry of Health and Welfare, and the Ministry of Land, Infrastructure and Transport, along with the data subcommittee chair of the National AI Strategy Committee, attended.
The first item on the agenda was the plan to fast-track the opening of the AI high-value public data TOP 100. The Ministry of the Interior and Safety has been releasing 100 categories of public data with high demand for AI development and significant social and economic ripple effects, and after consultations with the institutions that hold the data, it decided to bring the completion date forward to 2027. Under the previous plan, the data was to be opened in stages through 2028, but as the 10 categories released last year produced concrete results, the government resolved to supply quality data to domestic AI companies earlier. The revised schedule calls for 10 categories in 2025 (completed), 25 in 2026 (under way), and 65 in 2027; the volume originally split between 30 categories in 2027 and 35 in 2028 has now been consolidated into 2027.
The rationale for the early release lies in the results already demonstrated by data that has been opened. Synthetic data on transit-card usage released by the Korea Transportation Safety Authority was used to improve public transit systems and to analyze the operational effects of demand-responsive transport, while precedents of legal interpretation by central administrative agencies and rulings from special administrative appeals opened by the Ministry of Government Legislation served as training material for legal AI at legal-tech firms. The government said that from 2015 through this year it has opened 267 categories of high-quality public data, driving the creation of 3,625 private services, including the real-estate platform Zigbang, the hospital-booking service Goodoc, and the weather-information service Kweather.
The data to be released going forward broadly covers areas closely tied to daily life, including autonomous-driving simulation research data, synthetic data on taxi pickups and drop-offs, nationwide electric-vehicle charging data, real-time data on pediatric and obstetric emergency care, comprehensive data on infectious-disease management, and integrated traffic-accident investigation data.
The second agenda item was a plan to support public data and works for the advancement of independent AI models. To secure world-class AI competitiveness, the government has been pursuing an ‘independent AI foundation model’ project since August last year, and it has provided the elite teams involved with a total of 50 categories (about 23.93 million items) of public data and works managed by 33 ministries and agencies. These included historical materials from the National Institute of Korean History, reports from the Ministry of National Defense and the National Police Agency, and decisions from the Tax Tribunal.
Based on additional requests from the elite teams, the Ministry of Science and ICT and the Ministry of Culture, Sports and Tourism reviewed 102 categories of public data and works held by relevant institutions and decided to convert 37 categories free of copyright and personal-information concerns to the KOGL ‘AI type’ and similar licenses so that all AI companies can use them. The categories include agencies’ policy materials and publications, resolutions and rulings from various committees along with casebooks, and past questions from certain qualification examinations. The Korea Open Government License (KOGL) is the labeling standard for free use of public works; this January a new ‘AI type’ allowing free use for AI training was established. In addition to these 37 categories, each ministry will conduct its own review of how KOGL labels are attached to the public works under its jurisdiction and give priority to attaching the AI type, while the Ministry of Culture, Sports and Tourism will also pursue a revision of the Copyright Act to expand the opening of public works.
Ahead of the meeting, Prime Minister Han stressed, ‘Public data is an irreplaceable national asset accumulated over a long period across every economic and social field, and global AI competitiveness hinges on how actively quality public data is opened and how well it is used.’ She added, ‘Public data was built with the people’s precious tax money, so opening it is a natural principle,’ and urged the relevant ministries ‘to reflect on whether there have been areas where data was opened only passively, citing practices and other reasons.’
Closing the meeting, Prime Minister Han said, ‘We will spare no support so that AI and AX data can be well used not only in the public sector but also on the ground in the private sector,’ adding, ‘I will see to it responsibly that the measures discussed today lead to changes that people can feel in the field.’ In returning the public asset of nationally accumulated data into a resource that citizens and businesses can actually use, the measure carries significance in fostering the AI industry while broadening equity in data access.
지역 홍보대사 신청 기자양성과정 신청 코리안투데이 지부장 신청 코리안투데이 인천강화코리안브랜드대상 신청 최고경영인대상 신청 소비자만족도대상 신청
<저작권자 ⓒ 코리안투데이(The Korean Today) 무단전재 및 재배포 금지>
