토픽모델링(BERTopic)을 활용한 국내 인공지능 정책 관련 언론보도 분석

초록

This study analyzed Korean media coverage of AI policy after the rise of generative AI and examined the structure of policy-related discourses. A total of 12,572 news articles were collected from the BigKinds between November 2022 and July 2025, and BERTopic-based topic modeling was conducted. The analysis identified 15 topics, including ‘national digital strategy and policy governance’ and ‘local foundation model R&D and platform strategy.’ These topics were then grouped into five broader frames—strategy and governance, technological sovereignty and infrastructure, private sector and market ecosystem, risk and normative response, and international strategy and diplomacy—to examine framing patterns over time. The results show a sharp increase in AI policy-related coverage in 2025, with the technological sovereignty and infrastructure frame and the strategy and governance frame receiving the greatest emphasis. The findings also indicate that Korean AI policy discourse is jointly constructed by a diverse set of actors, including the government, local institutions, private firms, and international organizations, while government-led agendas still dominate news coverage. This study demonstrates that AI policy discourse in Korea is expanding beyond a technical focus toward processes of institutionalization, industrialization, and internationalization, and it highlights the role of the news media as a key intermediary in emerging AI governance.

키워드