The South Korean government plans to develop an AI-powered public security system by the end of 2026. According to reports from the Korea Times, Minister of Science and Information Technology Bae Kyung-hoon announced that his administration will begin developing a dedicated security intelligence model in the coming year. The first application of this technology will be in administrative offices later this year.

This initiative falls under efforts to build national capabilities in artificial intelligence and limit reliance on foreign companies providing models and infrastructure. The project should start using existing technologies available in the country. Minister Bae explained that current models could undergo adaptation using data related to security and vulnerability scenarios.

Based on the assessment illustrated by Bae, the current level of technological sophistication would allow for creating a system capable of identifying a wide range of vulnerabilities. The initial phase will focus on the public sector, designated as the starting environment for implementing the model. For businesses and other private operators, the government will evaluate the possibility of making the same technology available.

Minister Bae linked the intervention to international competition for so-called 'sovereign' AI, or the ability of a nation to develop and manage its own cutting-edge technologies. According to Bae, advanced models are increasingly being treated more like strategic resources than mere tools. In a reconstruction published by the Korea Times, Bae stated that the US is shifting towards promoting open-source models rather than closed systems.

Even China, currently associated with accessible models, may follow an analogous path. Minister Bae predicted a strengthening of the trend among governments to control frontier models, arriving at valuing them strategically akin to nuclear arms. The development of the public security system remains distinct from the goal of building a more expensive, comparable-to-Mythos-style sovereign AI model.

Bae clarified that no decision has been made yet. Remaining decisions include funding for any potential sovereign AI model, characteristics of the operational system, and conditions for adopting it later by the private sector.