On the 40th day of its pilot run, the "artificial intelligence (AI) government lab" has registered 405 experimental projects. That means an average of 10 experimental projects opened per day.
According to the Ministry of the Interior and Safety on the 12th, subscribers to the "AI government lab" public development artifact repository (GitLab), which opened on the 3rd of last month, surpassed 1,000. The "AI government lab" is a development and testing space where public officials use AI coding tools to directly implement field ideas. Developed code and experience can be shared through the public development artifact repository "GitLab." Of the current development projects, 189 are open on GitLab. An official at the Ministry of the Interior and Safety (MOIS) said, "It is a structure where other public officials can review the task details and development process, exchange opinions, or use them for follow-up development," adding, "A foundation is being laid for an individual public official's idea to lead to governmentwide sharing, collaboration, and diffusion."
The activities of "heavy users" who upload many experimental tasks are also drawing attention. Park Jeong-hwan, Director of the Medical Artificial Intelligence Data Policy Division at the Ministry of Health and Welfare, is pursuing 49 experimental tasks during this period. Kim Dong-hun, a researcher at the National Radio Research Agency, has 26, and Jeong Jun-u, Director of the Public Artificial Intelligence Innovation Division at the Ministry of the Interior and Safety, has also uploaded 21 experimental tasks.
Public officials at local governments are keeping pace. Park Seong-bok, Assistant Deputy Director at Gimhae City's AI Policy Division, is pursuing 16 experimental tasks, and Ryu Seung-in, Assistant Deputy Director at Seoul's Gwangjin District, is pursuing 13.
Project fields are also diverse. From report writing and proper language use to automating repetitive tasks and managing memos and to-dos, practical tools needed in actual administrative settings are being created.
Director Park Jeong-hwan is testing a knowledge management system that standardizes and accumulates division materials such as reports and minutes on the public GitLab, and has AI automatically classify and summarize them for use in material searches and drafting report outlines.
"OnAI Work," developed by Director Jeong Jun-u, drafts a public official's report when the task content is entered in natural language. It is being assessed as an AI tool that minimizes time spent on report writing and helps focus on policy decisions.
The Ministry of the Interior and Safety (MOIS) plans to prepare "guidelines for direct development by public officials," covering security, quality and deliverable management, and procedures for work utilization, so that public officials' AI-based development activities can proceed systematically. It will also support ways to apply outstanding projects validated in the lab to actual administrative work. If there is a need to develop them into formal public-facing services or institutionwide common services, the ministry plans to expand them in connection with informatization project procedures.
Hwang Gyu-cheol, head of the artificial intelligence government office at the Ministry of the Interior and Safety, said, "The biggest achievement of the AI government lab is that public officials who know the field best have begun to define problems and build solutions themselves, regardless of their affiliation or rank," adding, "We will support a single public official's small idea so it can lead to organizational work innovation and develop into a public asset used governmentwide."