Data ministry explores AI-friendly global data standards with OECD, ADB
Vice Data Minister Lee Myoung-ho, sixth from left, poses with other participants during an international workshop on artificial intelligence-driven innovation in data standards at a hotel in Seoul, Tuesday. Courtesy of Ministry of Data and Statistics
Korea is stepping up efforts to make official statistics and other public data more accessible to artificial intelligence (AI) and explore AI-friendly global data standards to develop data that AI can more accurately understand and use, the Ministry of Data and Statistics said Tuesday.
Toward that end, in collaboration with the OECD and the Asian Development Bank (ADB), the ministry opened a three-day workshop in Seoul, Tuesday, bringing together more than 80 data experts from statistical agencies and central banks in 25 countries.
A key focus is Statistical Data and Metadata eXchange (SDMx), an international standard that enables statistics and related information to be exchanged across different systems.
Various sessions cover ways to standardize and manage data for AI and use AI tools to improve data quality and streamline data production and management. Experts in data standards and AI from major international organizations, including the OECD and the Bank for International Settlements, serve as instructors.
On the final day, participants will take part in the SDMx + AI Innovation Hackathon, using the model context protocol to develop and present data service models that enable AI to access and use official statistics.
Minister of Data and Statistics Ahn Hyung-jun
“Through this workshop, the international statistical standard SDMx and AI were brought together in a practical setting, demonstrating how data and AI can enhance each other’s value,” Minister of Data and Statistics Ahn Hyung-jun said in an interview with The Korea Times.
“We will be able to strengthen practical capabilities to adopt global standards and cutting-edge technologies into the national data framework by sharing the latest technological trends and implementation experiences with international organizations and Asia-Pacific data and AI experts.”
Ahn noted that the ministry’s vision for data governance is to safely connect fragmented data at home and deliver its value to people’s lives, while working with the international community to turn Korea’s experience into global standards.
The minister also emphasized the importance of “good data,” which he defined as reliable, high-quality and well-structured data backed by rigorous quality management and standardization.
He vowed to prioritize developing AI-friendly statistical metadata to ensure that AI can accurately understand and use official statistics, pointing to a growing problem in which AI systems rely on secondary sources rather than official statistical databases, increasing the risk of inaccurate information.
The ministry therefore plans to build data infrastructure that allows AI to reliably search for, select and use official statistics while minimizing the risk of hallucinations.
"The establishment of AI-friendly statistical metadata presents an opportunity to chart a course for the metadata structures, standards and principles of the AI era in the international community," he said.
He stressed that the ministry will ensure international standards and compatibility in its metadata architecture and design while sharing its expertise and actively contributing to global discussions.
“Building upon these milestones, we will take a leading role in shaping global standards for the AI-ready statistical data ecosystem,” he said.