Methodology of North Korea Studies and Systematization of Knowledge through Artificial Intelligence

YIKUS Issue Brief No. 158

Kwon So Young (Senior Research Fellow, Yonsei Institute for Unification Studies, Yonsei University)

Research on North Korea has progressed steadily despite epistemological limitations and challenges in data collection. The methodologies used in North Korean studies have gradually become more specialized and diverse, moving beyond traditional perspectives. With the rise of the global era, extrinsic approaches based on universal human values or area studies became more popular than intrinsic approaches that emphasize North Korea’s uniqueness. Recently, analytical techniques that are scientific, systematic, and consistent have also been employed. Despite these efforts, understanding of North Korean society and the regime remains limited. As national interest in unification preparation increases, the demand for objective indicators and data reflecting North Korea’s current situation has grown. However, most research related to North Korea has been subjective due to restricted access to information caused by the regime’s closed nature and the characteristics of North Korean media, which limit the use of objective data. Quantitative research has depended on researchers’ interpretations of primary sources such as officially released newspapers and broadcasts. Since the 2000s, methods like interviews and surveys of North Korean defectors have been used to understand reality, but limitations remain because of problems with sample representativeness and the absence of verification tools. The lack of understanding of North Korean society and the regime is also clearly evident in domestic media reports, where speculative coverage of senior officials often turns out to be false, undermining the credibility of Korean media. It is often impossible to verify whether reports about purged officials returning are true or false, but more objective and quantitative evidence could have allowed for more comprehensive judgments.

The lack of systematic knowledge about North Korea is a major barrier to rational policy-making by stakeholders. Domestically, the absence of objective and structured information has intensified factional conflicts over North Korea-related issues. Therefore, it is essential to introduce quantitative research methods that organize knowledge and generate policy insights through objective analysis. One such method is the digitalization of North Korean studies, which combines artificial intelligence and data science techniques. This approach can deepen understanding of North Korean society and the regime by analyzing North Korean newspapers and broadcast reports, creating objective indicators, building searchable data engines, integrating these with other North Korea-related data, and ultimately developing a North Korean data ecosystem. In North Korean studies, where objective data are limited, media reports hold significant value as data sources. Particularly, with the rapid growth of big data and data science in social sciences, the value of unstructured data like media reports has increased. North Korean media should be viewed less as a “mirror” of reality and more as messages conveying the regime’s intentions to the world. Analyzing these reports in the context of North Korea’s environment can enhance understanding of internal conditions. In this light, it is advisable to supplement media reports with other hard data to generate key information that helps assess North Korea’s reality.

Despite challenging conditions, many foreign policy research institutions have recently made notable progress by expanding empirical data-based knowledge on North Korea. A notable example is the U.S. Stimson Center’s 38 North, which has attracted global media attention by analyzing commercial satellite images to observe changes in North Korea’s nuclear development. Since 2012, when 38 North started satellite image analysis, it has been cited 88 times in The New York Times and 333 times in The Washington Post. In South Korean media, it has been cited 15,490 times according to Naver News search results. This indicates a strong demand for information provided by foreign institutions conducting data-driven analysis. Conversely, South Korea, as the key stakeholder in unification, lags behind foreign institutions in providing insights on North Korean society and the regime. In this context, expanding quantitative research is crucial for systematically building knowledge about North Korea. The scope of North Korea research is broadening to include fields such as medicine, engineering, urban planning, and economics, while advances in big data technology offer new opportunities to overcome traditional research limitations. By adopting innovative analytical methods, it is important to expand interdisciplinary research and international scholarly collaboration, exemplified by digitally driven North Korean studies.

In South Korea, a pilot study was conducted by Professor Hahn Kyu Sup from the Department of Media and Communication at Seoul National University’s Institute for Peace and Unification Studies. Follow-up discussions were also held with George Mason University and the 38 North team in the U.S. Meetings with experts and think tanks in the U.S. emphasized the importance of Korea-led, data-driven North Korea research. First, generating objective data for policymaking can influence not only South Korea’s North Korea policy but also the policies of other stakeholders and international organizations, ensuring Korea’s position is properly represented. This should be promoted at the national level, rather than left solely to individual researchers. South Korea’s strengths in North Korea research include linguistic similarity, cultural understanding, and IT capabilities. With many high-quality researchers fluent in Korean language and as a global IT leader, pioneering digital North Korean studies based on these strengths is the best way for Korea to lead the globalization of North Korea research and policy studies. Meanwhile, foreign policy research institutions that shape discourse on North Korea, impacting U.S. government policy, tend to be weak in these areas. The number of Koreans actively working in top U.S. think tanks is limited, and recruiting Korean-speaking personnel for data training is challenging. Most researchers are policy or international relations experts lacking advanced data science skills or large-scale data collection experience. Even 38 North currently analyzes satellite images mainly through qualitative methods without incorporating engineering technologies like image recognition. By leveraging linguistic similarity and IT technology to generate data with policy implications and uniqueness, Korea can establish a distinctive North Korea research domain separate from foreign think tanks. Strengthening exchanges with U.S. universities and policy research centers, key stakeholders in North Korea-related issues, is vital. North Korea is interested in Washington think tanks influencing U.S. policy, and such networks can help South Korea take a leading role in dialogue with North Korea.

Second, providing data to global North Korea researchers and media is feasible. In the long run, a preliminary version of a database can be shared with university libraries worldwide, supplying valuable quantitative data to researchers. Through ongoing R&D, the database can be expanded and transformed into a North Korean media analysis engine capable of real-time text and video searches. This will enable researchers to use data flexibly and integrate it with other sources, supporting the systematic accumulation of North Korea-related knowledge. This approach will enhance South Korea’s role in North Korean studies and position it as a global leader. Currently, many media outlets produce speculative reports due to a lack of objective data. Developing and sharing an information platform for global journalists based on regular and objective data analysis will promote scientific and balanced reporting worldwide, broadening understanding of North Korea. However, accomplishing this requires engineering expertise and interdisciplinary collaboration with AI research institutes or engineering universities.

To overcome methodological exceptionalism, theoretical isolation, and ideological debates in North Korea research, diverse methodologies must be adopted to accurately reflect North Korea’s reality. If South Korean academia identifies research areas within North Korean studies where it holds comparative advantages and leads international efforts, it can amplify the voices of South Korean researchers in policy and academia and expand competitive research domains. The integration of North Korean studies and data science presents a historic opportunity for Korea to produce objective data for policymaking and lead global North Korea research beyond academia.

By Kwon So Young (Senior Research Fellow, Yonsei Institute for Unification Studies, Yonsei University)

● The Issue Brief is written based on the author’s views and has no relation to the official position of the Yonsei University Institute for Unification Studies or North Korean Review.