"AI Data International Standards, Led by Korea"

[Asia Economy Reporter Seulgina Jo] The South Korean delegation participating in the international standardization meeting on artificial intelligence (AI) technology led technical discussions across all areas including AI data, data quality, reliability, and industrial applications, laying the groundwork for South Korea to enter the ranks of leading countries in international standardization in this field.

The Ministry of Science and ICT’s National Radio Research Agency and the Ministry of Trade, Industry and Energy’s National Institute of Technology and Standards announced that at the ‘5th AI International Standardization Meeting (ISO/IEC JTC1/SC42)’ held until the 21st, the Korean delegation expanded the scope of work for AI data standardization, a core element of AI, and reflected South Korea’s cases in international standard documents, achieving significant results.

At this meeting, the research results of the AI Data Special Working Group, newly established last October under South Korea’s leadership, were shared, and it was agreed to expand the scope of data standardization work by the international standardization organization from big data standardization to “all data applicable to AI-related industries” for standardization going forward.

Accordingly, South Korea plans to propose a new project to derive AI data standards expanded to data standards for various industries utilizing AI, along with existing big data standards, by the next meeting in October this year, together with AI-leading countries such as the United States, China, Japan, and Canada.

Additionally, South Korea proposed the establishment of an Advisory Group to promote the standardization of the ‘AI service ecosystem’ at this meeting, planning to lead international standards across the entire industry.

Technical standards for data quality for machine learning were also proposed as a new project. Data quality refers to whether data has been processed through appropriate procedures and formats throughout the entire process of collection, processing, and distribution, and it is a key factor in enhancing AI performance. South Korea is leading international standardization by defining the concept and scope of standardization, with the United States and Germany planning to participate in work on quality verification procedures and measurement methods.

The international standard draft for the creation of a big data reference architecture, proposed by South Korea in December 2014, is expected to be published as a technical report at the end of this month following the conclusion of discussions at this meeting.

Along with this, South Korea added 17 more AI application cases to the technical report, following the inclusion of 8 cases last year. Furthermore, the need for new standardization work on ‘explainable artificial intelligence (XAI, eXplainable AI)’ to ensure the reliability and transparency of AI technology was also presented.

Kim Jeong-ryeol, Director of the National Radio Research Agency, evaluated, “By expanding the scope of international standardization work on AI through cooperation among industry, academia, research institutes, and government, we have achieved results that can serve as a foothold for our companies’ overseas market entry in the future, such as proposing new international standardization for our data-related technologies and promoting the establishment of an advisory group.”

Due to the impact of the novel coronavirus disease (COVID-19), this meeting was held online from the 6th to the 21st, with over 200 experts from more than 30 countries including the United States and Japan participating. The South Korean delegation consisted of 33 experts from industry, academia, research institutes, and government, including Professor Young-im Cho of Gachon University, Seong-jun Lim, Head of K-Data, Director Kang-chan Lee of ETRI, and Professor Jae-ho Lee of the University of Seoul.