- MSIT has prepared the draft Enforcement Decree by specifying and clarifying the matters delegated by the Artificial Intelligence Basic Act (“AI Basic Act”), based on extensive consultations with stakeholders.
- Public comments will be collected from November 12 to December 22, 2025.
- A grace period of at least one year will be implemented before imposing administrative fines, providing companies with sufficient preparation time and supporting them in understanding and fulfilling their obligations.
The Ministry of Science and ICT (MSIT, Deputy Prime Minister and Minister Bae Kyunghoon) announced that it will issue a legislative notice for the draft Enforcement Decree of the AI Basic Act (hereinafter the “Enforcement Decree”) for a 40-day period from November 12 to December 22, 2025. This legislative notice is being issued in line with the enactment of the Act on the Development of Artificial Intelligence and Establishment of Trust (“AI Basic Act”), which was promulgated on January 21, 2025, and is scheduled to take effect on January 22, 2026.
The AI Basic Act, enacted last December through bipartisan agreement to promote the AI industry and establish a foundation for AI safety and trust, is set to take effect early next year. MSIT prepared the draft Enforcement Decree based on extensive input from industry, civic groups, experts, and relevant ministries to ensure that the legislative intent of the Act is effectively reflected
MSIT presented the direction for preparing the subordinate statutes under the AI Basic Act to the National AI Strategy Committee on September 8 and released draft notices and guidelines on September 17 to gather public comments on all subordinate regulations. MSIT reviewed and incorporated the opinions received during this process and will continue to collect additional comments throughout the legislative notice period.
*Drafts of the Enforcement Decree, two notices, and five guidelines are publicly available on the website of the National Information Society Agency (NIA), and public comments on these documents will be accepted until January 21, 2026.
Taking into account global regulatory trends and the growth of the domestic AI industry, MSIT sought to introduce a flexible regulatory framework that places greater emphasis on promotion than restriction. Opinions from relevant ministries* were thoroughly reviewed to minimize overlapping or duplicative regulations.
*For example, when obligations under sector-specific laws—such as the Digital Medical Products Act— are fulfilled, the corresponding obligations for high-impact AI operators under the AI Basic Act will be deemed satisfied.
Reflecting these principles, the draft Enforcement Decree specifies and clarifies the matters delegated by the Act to support its effective implementation, ensuring a balance between fostering the AI industry and establishing a robust safety and trust framework. The major provisions are as follows.
1. Clarifying standards for support programs to foster the AI industry
To promote the domestic AI industry*, the Decree clearly defines the targets, criteria, and details of support programs prescribed by law—including AI R&D, training-data construction, and AI adoption and utilization—and establishes criteria and procedures for designating AI clusters, thereby providing a concrete institutional foundation to support the industry.
*Support areas include AI R&D, standardization, training-data construction, AI adoption and utilization, support for SMEs and startups, industrial convergence, talent development, international cooperation, overseas expansion, and data-center policy development.
2. Defining the designation and operation of supporting institutions for national AI policy
The Decree details the designation and operation of institutions that will drive AI industry development and support the establishment of a safety and trust framework. These include the AI Safety Institute (responsible for specialized AI safety functions), the AI Policy Center (for policy development and international regulatory cooperation), and a dedicated agency to support the operation of AI clusters.
3. Specifying and clarifying systems to ensure AI safety and trust
To reduce business uncertainty, the Decree elaborates on the statutory requirements concerning transparency and safety obligations, criteria for determining high-impact AI, and AI impact assessments.
For transparency, providers of products or services using high-impact or generative AI must notify users in advance that AI is being used. When an output is difficult to distinguish from reality, users must be clearly informed that it was generated by AI, taking into account factors such as age and physical condition.
For safety obligations, AI systems trained with cumulative compute of at least 10²⁶ floating-point operations (FLOPs) are designated as subject to safety requirements, reflecting international standards and developments in AI technology.
High-impact AI is assessed according to detailed criteria—such as application area, risks to basic rights, severity, and frequency—and a specified confirmation procedure. MSIT’s confirmation process requires 30 days, extendable once for up to an additional 30 days, with written notice of the reasons and duration of the extension to reduce business burdens.
The Decree also sets out the minimum requirements for AI impact assessments, including which fundamental rights are affected, how they are affected, and potential mitigation measures, ensuring that the system operates effectively.
To support early implementation of the safety and trust framework, MSIT will additionally prepare detailed notices and guidelines outlining specific standards, compliance methods, and best practices to prevent unintended misperceptions that the regulatory scope is being expanded.
To support the effective implementation of the Act in its early stage and provide companies with sufficient preparation time, MSIT will implement a grace period of at least one year before administrative fines are imposed. Efforts are currently underway to gather opinions to finalize the detailed operation plan and duration of this grace period.
During the grace period, MSIT will operate an integrated support platform for the AI Basic Act through a dedicated center, tentatively named the Integrated Guidance Support Center. The center will provide detailed guidance to help companies address difficulties related to the application of the Act and other legal requirements.
In addition to providing guidance, the platform will function as a venue for continuously collecting feedback on the Enforcement Decree and related guidelines. The insights gathered will be reflected in future revisions, enabling the regulatory framework to evolve in line with the rapidly changing AI environment.
Financial support measures will complement these efforts. After the Act takes effect, MSIT will secure funding to cover the costs of AI testing and certification and AI impact assessments, while continuing to refine guidelines. MSIT will also seek expert advice on transparency obligations and the responsibilities of high-impact AI operators.
Minister Bae stated, “The draft Enforcement Decree of the AI Basic Act will serve as a critical institutional foundation for Korea to secure its position as one of the world’s top three AI powerhouses.” He added, “During the legislative notice period, we will gather a wide range of field opinions to ensure that the Enforcement Decree fully reflects the Act’s purpose—promoting the AI industry and establishing a foundation for safety and trust.”
For further information, please contact the Public Relations Division (Phone: +82-44-202-4034, E-mail: msitmedia@korea.kr) of the Ministry of Science and ICT.
Please refer to the attached PDF.