As South Korea accelerates its national AI strategy, the government is preparing to launch a free, homegrown chatbot service that will serve as public digital infrastructure akin to schools or libraries. The initiative is framed as a cornerstone of long‑term digital sovereignty, embedding AI access into the basic services the state guarantees to residents. The service is explicitly intended to help underserved populations access AI tools and public information, narrowing the country’s digital divide. Managers will play a critical role in workflow redesign, ensuring effective integration of this chatbot into various public services.
Officials describe the program as a way to reduce reliance on foreign AI platforms such as ChatGPT and Claude by prioritizing Korean foundation models in government and citizen services. The national AI foundation model has been designated the default for public‑sector transformation projects, and ministries must seek explicit approval if they choose foreign systems instead. By channeling demand toward domestic providers, the government aims to strengthen the local AI ecosystem, build exportable capabilities, and guarantee that language, data, and policy preferences are reflected in the models citizens use.
The nationwide service, tentatively branded as “AI for All” or “AI for Everyone,” is positioned as a Korean‑style ChatGPT available to every resident without fees or usage caps. The Ministry of Science and ICT opened competitive bidding in mid‑July 2026 to select two or three operators that will build and run both the general‑purpose chatbot and a public‑service AI agent. A public beta of the chatbot is scheduled for late September 2026, followed by a full national launch before year‑end, aligning the rollout with broader AI policy briefings and implementation of new AI legislation.
More advanced agents are expected to be phased in through 2026 and 2027, culminating in a vision of “one AI agent per person.” Technical rules require that at least half of the core system run on Korean‑developed foundation models, with an additional share mandated to use other domestic AI technologies so that multiple vendors participate. Foreign models may supplement the stack, but their share is excluded from state support, reinforcing the priority on homegrown systems.
To meet the computing demands of a free, unlimited national service, the government is offering up to 512 Nvidia B200 graphics processing units to selected operators, alongside obligations for matching private investment. This architecture is designed for scalable, high‑volume use so that every resident can access AI tools without worrying about quotas or subscription costs.
The core offering will be a general‑purpose chatbot for everyday tasks such as information search, writing assistance, learning support, and productivity workflows, delivered through a single national interface. A secondary public‑service agent will help users discover welfare and administrative programs, check eligibility, and complete applications, functioning as a front door to the bureaucracy. The agent is expected to proactively identify benefits citizens may have overlooked, notify them, and execute the necessary steps on their behalf.
Over time, the roadmap includes specialized agents for education, personal finance management, retirement planning, and business support, gradually evolving into personalized assistants capable of handling reservations and complex life‑planning decisions. Together, these measures anchor domestic AI as everyday national infrastructure, nationwide.
