The United Korean AI Act Bill: Contents in Comparison with the EU AI Act (with English Translation of the Bill)

In a landmark development for AI governance in South Korea, the National Assembly has introduced a unified bill known as the AI Basic Act. This legislation consolidates nineteen separate AI-related proposals submitted between May and November 2024. The Science, Technology, Information, Broadcasting and Communications Committee spearheaded the merger of these proposals into a comprehensive alternative bill. While the bill still requires approval from the Judiciary Committee and a final plenary session, its core principles are expected to remain intact, signaling a transformative step toward establishing a robust legal framework for AI development, promotion, and regulation.

Comparative Analysis of South Korea’s AI Act bill and the European Union’s AI Act

Below is a comparative analysis of the European Union’s AI Act and South Korea’s AI Act bill. While both aim at establishing a robust regulatory framework for artificial intelligence, their approaches, priorities, and institutional contexts reveal both convergences and divergences.

1. Scope and Objectives

• EU AI Act: Article 1 states that the Act aims to ensure that AI systems placed on the EU market respect existing law, fundamental rights, and Union values. It sets harmonized rules for the development, placing on the market, and use of AI systems.
• Korean Bill: Article 1 (Purpose) clearly mirrors the spirit of the EU’s objective. It establishes that the law’s goal is to contribute to the protection of rights and dignity, improve quality of life, and strengthen national competitiveness by setting basic rules for AI’s sound development and the establishment of a trust-based foundation.

Comparison: Both frameworks begin by emphasizing fundamental rights and societal interests rather than mere economic objectives. While the EU references “Union values” and fundamental rights directly, the Korean bill uses language like “rights and dignity” and “trust-based foundation,” showing conceptual alignment.

2. Definitions and Scope of “High-Risk” or “High-Impact” AI

• EU AI Act: Article 6 defines “high-risk AI systems” based on their intended purpose and the product safety legislation listed in Annex II and Annex III. For example, Annex III includes AI used in critical infrastructure, education, employment, essential public services, and law enforcement.
• Korean Draft: Article 2(4) defines “high-impact AI” as AI systems affecting “life, bodily safety, and fundamental rights” and provides a list of domains (e.g., energy supply, healthcare, medical devices, public services) which substantially overlap with the EU’s Annex III categories. For instance, Korean draft Article 2(4)(c) deals with healthcare, comparable to the EU’s classification of AI in medical devices as high-risk.

Comparison: Both laws classify certain AI applications as inherently more sensitive. The EU uses “high-risk” while Korea uses “high-impact,” but both hinge on the societal importance of these sectors and the potential harm to individuals. The substantive areas largely match: medical, safety-critical services, public services, and fundamental rights–sensitive uses.

3. Transparency and User Notification

• EU AI Act: Article 52 requires that users be informed when they are interacting with an AI system that may present a risk or when AI-generated content is used, especially in “deepfakes.” For instance, Article 52(3) mandates disclosure of AI-generated or manipulated content that could deceive the public.
• Korean Draft: Article 31 (AI Transparency Obligations) obliges AI business operators to inform users in advance if a product or service uses high-impact or generative AI. Article 31(2) demands that results generated by generative AI be clearly labeled, and Article 31(3) addresses synthetic media (“genuine-looking” AI outputs), requiring that users be clearly notified.

Comparison: Both legal texts impose transparency obligations. The EU Act’s Article 52 and the Korean draft’s Article 31 are functionally similar in mandating that users must know when they deal with AI-generated content. The Korean draft’s mention of “generative AI” parallels the EU’s concern about deepfakes and manipulated media.

4. Risk Management and Technical Requirements for High-Risk/High-Impact AI

• EU AI Act: Articles 9 to 15 detail obligations for high-risk AI systems, including risk management systems (Article 9), data governance and data quality requirements (Article 10), and technical documentation (Article 11). These include continuous risk assessment, post-market monitoring (Article 61), and ensuring that the system meets conformity assessments before being put on the market.
• Korean Draft: Article 32 (AI Safety Obligations) requires AI business operators exceeding a certain computational threshold to implement risk identification, assessment, and mitigation. Article 34 (Responsibilities for High-Impact AI) sets forth measures similar to the EU’s—requiring risk management plans, explanation capabilities, user protection measures, and human oversight. Additionally, Article 35 (Influence Assessment) encourages impact assessments on fundamental rights.

Comparison: The EU imposes a rigorous risk management framework for high-risk AI. Korea, in Articles 32 and 34, similarly demands risk management, documentation, and oversight for high-impact AI. While the EU Act specifies a structured, pre-market conformity assessment (Articles 43-51), the Korean draft implies post-hoc oversight and possible support from institutions like the AI Safety Research Institute (Article 12). The idea of an “impact assessment” (EU’s term) aligns closely with Korea’s “influence assessment” (Article 35), showing a shared regulatory philosophy.

5. Governance Bodies and Institutions

• EU AI Act: Articles 56-58 establish a European Artificial Intelligence Board to facilitate a harmonized application of the regulations. National competent authorities are designated for implementation and enforcement.
• Korean Draft: Article 7 creates a Presidential-level National AI Committee to coordinate AI policy, and Articles 11 and 12 provide for an AI Policy Center and an AI Safety Research Institute. These bodies work together to set guidelines, ensure compliance, and support policy development and international cooperation.

Comparison: Both frameworks rely on specialized bodies to guide implementation. The EU uses a Board to coordinate among member states, reflecting the EU’s multi-jurisdictional structure. Korea’s centralized approach (National AI Committee) and specialized institutes mirror the EU’s model of oversight and expertise but adapted to a unitary state context.

6. Ethical and Human Rights Foundations

• EU AI Act: While not always explicitly naming “ethics,” the entire structure is grounded in the EU’s fundamental rights framework. The Recitals and many Articles emphasize compliance with Union values, including non-discrimination (Articles 5, 10), safety, and respect for privacy (Article 10 on data quality includes preventing discriminatory outcomes).
• Korean Draft: Article 27 (AI Ethics Principles) allows the government to issue ethics principles and mandates the Minister of Science and ICT to develop implementation plans. This codifies ethics guidelines at a statutory level. Further, Article 3 (Basic Principles) and references throughout highlight dignity, rights, and prosperity of citizens.

Comparison: The EU Act’s human rights alignment is implicit through compliance with the EU Charter of Fundamental Rights and GDPR. Korea’s draft explicitly enshrines AI ethics principles, requiring the government to publicize them. Both aim at ensuring AI supports, rather than undermines, fundamental rights, but Korea’s approach is more explicit in legislating the development of ethical guidelines (Article 27), while the EU relies heavily on its existing human rights framework and references in the Act’s recitals and cross-references to other EU laws.

7. Enforcement and Penalties

• EU AI Act: Articles 59-71 detail enforcement, market surveillance, and penalties, including fines of up to 30 million euros or 6% of global annual turnover for non-compliance (Article 71).
• Korean Draft: Article 40 (Fact-Finding Investigations) grants the Minister of Science and ICT investigative powers. Articles 42 and 43 set forth penalties and administrative fines (fines up to 30 million won for certain violations). While penalties exist, the Korean draft’s penalty structure appears somewhat less extensive than the EU’s, which has a tiered and highly publicized penalty regime.

Comparison: Both laws have enforcement mechanisms and penalties. The EU’s system is more explicitly tied to global turnover and is potentially harsher. Korea’s approach, while including imprisonment and fines, is less severe in scale. Nonetheless, both allow authorities to order corrective measures and ensure compliance.

8. Innovation, Competitiveness, and International Dimension

• EU AI Act: The Act aims at balancing innovation with fundamental rights protection. While it places stringent obligations on high-risk AI, it does not delve deeply into supporting ecosystems or clusters directly within the Act. Most supportive measures come from other EU programs (Horizon Europe, Digital Europe Programme).
• Korean Draft: Articles 16-23 show a strong emphasis on nurturing AI industry clusters, supporting SMEs, fostering international cooperation (Article 22), building data infrastructure (Article 15), and promoting R&D and training. This indicates that Korea’s law incorporates industrial policy and capacity-building measures directly into the regulatory framework.

Comparison: Both seek to foster trustworthy AI markets, but the Korean draft integrates innovation and competitiveness measures within the same legal instrument that sets regulatory standards. The EU separates regulatory and promotional activities into different legislative and funding programs. Korea’s single act thus blends EU-style protective rules with proactive industrial development strategies.

Conclusion

When comparing the EU’s AI Act and Korea’s AI Act bill, we see a clear thematic convergence:

• Both identify certain areas where AI poses heightened risks (EU Articles 6, Annex III vs. Korean Article 2(4)).
• Both mandate transparency and user awareness (EU Article 52 vs. Korean Article 31).
• Both require robust risk management, documentation, and oversight for high-risk/high-impact AI (EU Articles 9-15 vs. Korean Articles 32, 34).
• Both are founded on principles of human rights, dignity, and trust, although the EU relies on existing EU values and Korea explicitly calls for AI ethics guidelines (EU Recitals and cross-references vs. Korean Articles 3, 27).
• Both establish oversight bodies and enforcement mechanisms (EU Articles 56-58 vs. Korean Articles 7, 11, 12, and enforcement in 40-43).

The main differences lie in the explicit integration of industrial support measures in the Korean bill (Articles 16-23) and the EU’s heavier reliance on separate initiatives. However, on core regulatory principles and approaches, the EU Act and the Korean bill show strong parallels, reinforcing the notion that Korea’s law draws inspiration from the EU’s regulatory framework while tailoring it to national objectives.

Basic Bill (Alternative) on the Development of Artificial Intelligence and the Establishment of a Trust-Based Foundation (English Translation)

1. Background of the Alternative Proposal

Date of Proposal: November 2024 Proposed by: Chairperson of the Science, Technology, Information, Broadcasting, and Communications Committee

At the 3rd meeting of the Information and Communications Broadcasting Bill Review Subcommittee of the 418th National Assembly (Regular Session) on November 21, 2024, the above 19 bills were merged and reviewed, resulting in a unified and coordinated version, and it was decided to propose an alternative by the Committee.

On November 26, 2024, at the 17th meeting of the Science, Technology, Information, Broadcasting, and Communications Committee of the 418th National Assembly (Regular Session), it was resolved, in accordance with the Subcommittee’s report, that none of the 19 bills would be referred to the plenary session individually, and that the alternative drafted by the Subcommittee would be proposed as the Committee’s bill.

2. Reasons for Proposing the Alternative

Artificial Intelligence (AI), as a foundational technology that transforms not only all industries but also society itself, is rapidly expanding its use in all areas of society due to its remarkable performance improvements. In particular, the emergence of generative AI has drawn worldwide attention not only to the potential benefits that AI may bring but also to concerns about the associated risks. To support the sound development of AI and establish a trust-based foundation for an AI-driven society, it is necessary to stipulate fundamental matters. This will protect the rights and dignity of the people, improve the quality of life, and strengthen national competitiveness, thereby setting a new standard for AI in the Republic of Korea.

3. Main Provisions

a. The purpose of this Act is to contribute to the protection of the rights and dignity of citizens, the improvement of quality of life, and the strengthening of national competitiveness by stipulating matters necessary for the sound development of AI and the establishment of a trust-based foundation (Article 1).

b. Terms such as AI, high-impact AI, generative AI, AI ethics, AI business operators, etc. are defined (Article 2).

c. The Minister of Science and ICT shall every three years develop and implement a Basic AI Plan, subject to deliberation and resolution by the National AI Committee, to promote AI technology and the AI industry and to enhance national competitiveness. The Basic Plan shall include matters on the basic direction of AI policy, the training of professionals, and the establishment of a trust-based foundation (Article 6).

d. A National AI Committee under the President shall be established to deliberate and resolve major policies on promoting the AI industry and establishing a trust-based foundation. The Committee shall deliberate and resolve on matters such as formulating the Basic Plan, facilitating AI utilization, and regulating high-impact AI (Articles 7 and 8).

e. The Minister of Science and ICT may designate an AI Policy Center to develop AI-related policies and establish and disseminate international norms, and may operate an AI Safety Research Institute to ensure AI safety (Articles 11 and 12).

f. The government may support projects related to the investigation of domestic and international trends and systems, commercialization, and research and development of AI technology. The Minister of Science and ICT may carry out projects for standardization of AI technology (Articles 13 and 14).

g. The Minister of Science and ICT may nurture professional personnel related to AI technology and promote measures to secure foreign professionals (Article 21).

h. The State and local governments may pursue the functional, physical, and regional clustering of enterprises, institutions, or organizations engaged in the research and development of AI and AI technology to enhance the competitiveness of AI development and utilization (Article 23).

i. The government may enact and publish an AI Ethics Principle, including safety, reliability, accessibility, and contributions to human life and prosperity, to spread AI ethics. The Minister of Science and ICT shall develop implementation plans for the AI Ethics Principle and publicize and educate on them (Article 27).

j. The Minister of Science and ICT may promote projects to support voluntary verification and certification activities for ensuring AI safety and reliability (Article 30).

k. AI business operators providing products or services using high-impact AI or generative AI shall notify users in advance of that fact. When providing generative AI products or services, it shall be indicated that the result was generated by generative AI. Furthermore, when providing AI-generated outputs that are difficult to distinguish from reality, such fact shall be clearly disclosed to the user (Article 31).

l. AI business operators shall ensure the safety of AI systems that surpass a certain threshold of cumulative computational usage for training by carrying out risk identification, assessment, and mitigation measures (Article 32).

m. AI business operators providing high-impact AI or products and services using such AI shall implement measures to ensure safety and reliability (Article 34).

n. If the Minister of Science and ICT discovers or suspects violations of this Act, they may require the submission of data from AI business operators or conduct necessary investigations. If a violation is confirmed, they may order the cessation or correction of the violation (Article 40).

Bill No. [ ] Basic Bill on the Development of Artificial Intelligence and the Establishment of a Trust-Based Foundation, etc.

Chapter 1 General Provisions

Article 1 (Purpose) The purpose of this Act is to contribute to the protection of the people’s rights and dignity, the improvement of the quality of life, and the strengthening of national competitiveness by stipulating fundamental matters necessary for the sound development of artificial intelligence (AI) and the establishment of a trust-based foundation.

Article 2 (Definitions) The terms used in this Act are defined as follows:

1. “Artificial intelligence” means an electronic implementation of intellectual capabilities that humans possess, such as learning, reasoning, perception, judgment, and language comprehension.
2. “AI system” means an AI-based system that, with various levels of autonomy and adaptability, infers predictive, advisory, or decision-making outputs for given goals to influence real or virtual environments.
3. “AI technology” means hardware, software, or related application technologies necessary to implement AI.
4. “High-impact AI” means an AI system that may have significant effects on or pose risks to human life, bodily safety, or fundamental rights. It specifically refers to AI systems utilized in any of the following areas: a. Supply of energy as defined in Article 2(1) of the Energy Act b. Production processes of drinking water as defined in Article 3(1) of the Drinking Water Management Act c. Provision and utilization system of healthcare as defined in Article 3(1) of the Framework Act on Health and Medical Services d. Development and use of medical devices as defined in Article 2(1) of the Medical Devices Act and digital medical devices as defined in Article 2(2) of the Digital Medical Products Act e. Safe management and operation of nuclear materials and nuclear facilities as defined in Article 2(1)1 and 2 of the Act on Physical Protection and Radiological Emergency f. Analysis and utilization of biometric information (facial, fingerprint, iris, palm vein, etc.) used for criminal investigation or arrest g. Decision-making or evaluation significantly affecting personal rights or obligations such as recruitment or loan screening h. Major operations and management of transportation means, facilities, and systems as defined in Article 2(1)-(3) of the Transportation Safety Act i. Decision-making by the State, local governments, and public institutions (as defined in Article 4 of the Act on the Management of Public Institutions) that affects the public, including qualification verification, determination, or cost collection for public services j. Student evaluations in early childhood, elementary, and secondary education as defined in Article 9(1) of the Framework Act on Education k. Other areas specified by Presidential Decree that significantly affect human life, bodily safety, and fundamental rights.
5. “Generative AI” means an AI system that creates written, auditory, visual, or other diverse outputs by imitating the structure and characteristics of input data (as defined in Article 2(1) of the Framework Act on Data Industry Promotion and Utilization).
6. “AI industry” means an industry that develops, manufactures, produces, or distributes AI or AI technology-based products (“AI products”) or provides related services (“AI services”).
7. “AI business operator” means any corporation, organization, individual, or national institution engaged in AI-related business activities, falling under any of the following: a. AI developer: A person who develops and provides AI b. AI user business: A person who provides AI-based products or services utilizing AI developed by others
8. “User” means a person who is provided with AI products or AI services.
9. “Impacted individual” means a person whose life, bodily safety, or fundamental rights are significantly affected by AI products or services.
10. “AI society” means a society that generates value and drives development in all fields—industry, economy, society, culture, and administration—through AI.
11. “AI ethics” means the ethical standards that all members of society must observe in all areas related to the development, provision, and use of AI, based on respect for human dignity and aimed at protecting people’s rights, lives, and property to realize a safe and trustworthy AI society.

Article 3 (Basic Principles) (1) AI technology and the AI industry shall be developed to enhance safety and reliability, thereby improving the quality of life for citizens. (2) The State and local governments shall respect the creativity of AI business operators and strive to create an environment conducive to safe AI use. (3) The State and local governments shall devise policies so that all citizens can adapt stably to the changes that AI brings to society, economy, culture, and everyday life. (4) Where AI has a significant impact on human life, bodily safety, or fundamental rights, impacted individuals shall be provided, within technically and reasonably possible limits, with a clear and meaningful explanation regarding the reasons and principles behind the results produced by AI.

Article 4 (Scope of Application) (1) This Act shall apply to any acts conducted abroad that affect the domestic market or users in the Republic of Korea. (2) This Act shall not apply to AI developed or used solely for national defense or national security purposes as prescribed by Presidential Decree.

Article 5 (Relationship with Other Laws) (1) Except as otherwise provided by other laws, matters concerning AI, AI technology, the AI industry, and AI society (hereinafter “AI, etc.”) shall be governed by this Act. (2) When enacting or amending other laws related to AI, etc., consistency with the purpose of this Act shall be ensured.

Chapter 2 Institutional Framework for Sound AI Development and Trust-Based Foundation

Article 6 (Formulation of a Basic AI Plan) (1) The Minister of Science and ICT shall, every three years, formulate and implement a Basic AI Plan (hereinafter “Basic Plan”) to promote AI technology and the AI industry and strengthen national competitiveness, subject to deliberation and resolution by the National AI Committee established under Article 7, after seeking opinions from heads of relevant central administrative agencies and local governments. Minor amendments prescribed by Presidential Decree are exempt from this requirement. (2) The Basic Plan shall include the following matters:

1. Basic direction and strategies for AI policy
2. Training of professionals, building a trust-based foundation, and measures for systematic promotion of the AI industry
3. Legal, institutional, and cultural measures for establishing sound AI ethics and implementing an AI society
4. Securing resources and directions for investment in R&D and industrial promotion of AI technology
5. Ensuring fairness, transparency, responsibility, and safety of AI and measures to build trust
6. Projecting future developments of AI technology and corresponding measures in fields such as education, labor, economy, and culture
7. Other matters deemed necessary by the Minister of Science and ICT for promoting AI technology, the AI industry, international cooperation, and strengthening national competitiveness (3) The Minister of Science and ICT shall consider the comprehensive plans and execution plans under Articles 6(1) and 7(1) of the Intelligent Informatization Act when formulating the Basic Plan. (4) The Minister of Science and ICT may request data from other central administrative agencies, local governments, and public institutions for formulating the Basic Plan. Unless there are special circumstances, those agencies shall comply. (5) The Basic Plan shall be considered as the sector-specific implementation plan for AI and the AI industry under Article 13(1) of the Intelligent Informatization Act. (6) Heads of central administrative agencies and local governments shall consider the Basic Plan when formulating and executing their major policies. (7) Matters necessary for formulating and executing the Basic Plan shall be prescribed by Presidential Decree.

Article 7 (National AI Committee) (1) The National AI Committee (hereinafter “the Committee”) shall be established under the President to deliberate and resolve major policies on promoting the AI industry and establishing a trust-based foundation. (2) The Committee shall consist of no more than 45 members, including one chairperson and one vice-chairperson. A majority of the Committee members shall be civilians as defined in Paragraph (4)4, and it shall not be composed of only one gender. (3) The President shall serve as the Committee’s chairperson, and the vice-chairperson shall be appointed by the President from among the members falling under Paragraph (4)4. (4) The Committee’s members are as follows:

1. Government members: Heads of central administrative agencies as prescribed by Presidential Decree
2. The Vice Chief in charge of AI at the National Security Office
3. The Senior Secretary at the Presidential Office who assists in AI affairs
4. Civilian members: Persons with abundant expertise and experience in AI who are appointed by the President (5) The chairperson represents the Committee and oversees its affairs. (6) The chairperson may have the vice-chairperson act on their behalf as necessary. (7) The term of civilian members in Paragraph (4)4 is two years, renewable once. (8) The Committee may issue recommendations or opinions to national institutions, AI business operators, etc. regarding the proper use of AI, the implementation of AI ethics, and ensuring AI safety and reliability. (9) When the Committee issues a recommendation or expresses an opinion to a national institution on legislative or institutional improvement or action plans, that institution shall formulate corresponding action plans and implement them. (10) The Committee shall have one secretary member, who is the member under Paragraph (4)3. (11) Committee members shall not disclose confidential information learned in the course of their duties nor use it for any purpose other than their official duties, except as otherwise provided by law. (12) A support team shall be established under the Committee to assist its work. (13) The Committee shall remain in existence for five years from the date this Act takes effect. (14) Matters concerning the composition and operation of the Committee and its support team shall be prescribed by Presidential Decree.

Article 8 (Functions of the Committee) The Committee shall deliberate and resolve the following matters:

1. Formulation of the Basic Plan and checking and analyzing its implementation status
2. Policies related to AI, etc.
3. Strategies for research and development in the AI field
4. Investment strategies related to AI
5. Identification and improvement of regulations that hinder AI industry development and competitiveness
6. Plans for expanding infrastructure such as data centers as defined in Article 40(1) of the Intelligent Informatization Act
7. Facilitating AI utilization in manufacturing, services, and the public sector
8. International cooperation regarding the establishment of AI-related international norms
9. Recommendations or expression of opinions under Article 7(8)
10. Regulation of high-impact AI
11. Social changes related to high-impact AI and corresponding policy responses
12. Matters for which this Act or other laws require the Committee’s deliberation
13. Other matters that the chairperson deems necessary to submit to the Committee’s meeting

Article 9 (Disqualification, Challenge, and Recusal of Members) (1) A Committee member shall be disqualified from the deliberation and resolution of a specific agenda in the following cases to ensure fairness:

1. If the member or the entity to which the member belongs has a direct interest in the agenda
2. If the member’s family (as defined in Article 779 of the Civil Act) is an interested party in the agenda (2) Any party to the deliberation (including executives and employees of corporations or organizations) may file a challenge if there are circumstances where it is difficult to expect a fair performance of duties by a Committee member. If such a challenge is deemed valid, the Committee shall decide to exclude that member by resolution. (3) If a member falls under Paragraph (1) or (2), he or she shall voluntarily recuse themselves from the deliberation.

Article 10 (Subcommittees, Special Committees, and Advisory Panels) (1) The Committee may establish subcommittees by field of expertise as necessary. (2) The Committee may establish special committees as necessary to address particular AI-related issues. (3) The Committee may form advisory panels of experts for professional review of AI-related matters. (4) Matters necessary for the composition and operation of subcommittees, special committees, and advisory panels shall be prescribed by Presidential Decree.

Article 11 (AI Policy Center) (1) The Minister of Science and ICT may designate an AI Policy Center (hereinafter “Center”) to perform tasks necessary for the development of AI policy and the establishment and dissemination of international norms. (2) The Center shall carry out the following tasks:

1. Technical support necessary for formulating and implementing the Basic Plan
2. Technical support for the development and implementation of AI-related policies and projects
3. Investigation and analysis of how AI affects society, economy, culture, and people’s daily lives
4. Investigation and research of laws, systems, and trends to support AI-related policy development
5. Tasks entrusted by other laws or delegated by relevant agencies
6. Other tasks entrusted by national institutions (3) Matters necessary for the establishment and operation of the Center shall be prescribed by Presidential Decree.

Article 12 (AI Safety Research Institute) (1) The Minister of Science and ICT may operate an AI Safety Research Institute (hereinafter “Safety Institute”) to professionally and efficiently undertake activities to secure AI safety, i.e., protecting people’s lives, bodies, property, etc. from risks associated with AI and maintaining trust in an AI society. (2) The Safety Institute shall perform the following tasks:

1. Defining and analyzing AI-related risks
2. Research on AI safety policy
3. Research on AI safety evaluation criteria and methods
4. Research on AI safety technologies and standardization
5. International exchange and cooperation regarding AI safety
6. Support for securing the safety of AI systems as per Article 32
7. Other tasks related to AI safety as prescribed by Presidential Decree (3) The government may provide funding from the budget as necessary for the operation of the Safety Institute and the implementation of its projects. (4) Matters necessary for the operation of the Safety Institute shall be prescribed by Presidential Decree.

Chapter 3 AI Technology Development and Industrial Promotion Section 1 AI Industry Infrastructure

Article 13 (Support for AI Technology Development and Safe Use) (1) The government may support the following projects to activate AI technology development:

1. Investigation of domestic and foreign AI technology trends, levels, and related systems
2. Research, development, testing, evaluation, and utilization of AI technology
3. Dissemination, transfer, and commercialization of AI technology
4. Smooth distribution of information and industry-academia cooperation for implementing AI technology
5. Other businesses related to AI technology development and research as prescribed by Presidential Decree (2) The government may support the following projects to ensure the safe and convenient use of AI technology:
6. Research and development projects related to implementing the matters of Article 60(1) of the Intelligent Informatization Act in AI
7. Support for research and dissemination of technologies that implement the emergency stop function of Article 60(3) of the Intelligent Informatization Act into AI
8. Research, development, and dissemination of design standards and technologies suitable for privacy protection under Article 61(2) of the Intelligent Informatization Act
9. Research and development projects to conduct and apply social impact assessments under Article 56(1) of the Intelligent Informatization Act to AI
10. Research and dissemination of technologies or standards to ensure that AI is developed and used to respect human dignity and fundamental rights
11. Projects for improving awareness of safe AI development and use, and for educating and publicizing proper usage methods and safe environments
12. Other projects necessary to protect fundamental rights, physical and property safety in AI development and use (3) The government shall make the results of projects under Paragraph (2) easily accessible to anyone. Where necessary to protect developers, a protection period may be designated during which technology usage fees may be collected or other protective measures may be taken.

Article 14 (Standardization of AI Technology) (1) The Minister of Science and ICT may pursue the following projects for standardizing AI technology and related areas, such as training data, AI safety, and reliability. However, if Korean Industrial Standards under the Industrial Standardization Act are already established, those standards shall apply. If it is necessary to enact, revise, or abolish Korean Industrial Standards, it shall be done in accordance with the Industrial Standardization Act.

1. Establishing, revising, and abolishing standards related to AI technology and their dissemination
2. Investigating and researching domestic and international standards for AI technology
3. Other standardization projects related to AI technology (2) The Minister of Science and ICT may announce the standards established under Paragraph (1)1 and recommend related business operators to comply with them. (3) The Minister of Science and ICT may provide necessary support for private sector standardization activities related to AI technology. (4) The Minister of Science and ICT shall maintain and strengthen cooperation systems with international standardization organizations or bodies concerning AI technology standards. (5) Matters necessary for the promotion and support of standardization projects shall be prescribed by Presidential Decree.

Article 15 (Measures for AI Training Data, etc.) (1) The Minister of Science and ICT shall, in consultation with relevant central administrative agencies, promote necessary measures to facilitate the production, collection, management, distribution, and utilization of data used for AI development and utilization (“training data”). (2) The government may select and support target projects to facilitate training data production, collection, management, distribution, and utilization. (3) The government may implement training data construction projects to produce and provide various training data to stimulate the production, collection, management, distribution, and utilization of training data. (4) The Minister of Science and ICT shall establish and manage an integrated provision system (“integrated provision system”) that can comprehensively provide and manage training data from training data construction projects and allow free use by the private sector. (5) The Minister of Science and ICT may charge fees to users of the integrated provision system. (6) Matters necessary for selecting and supporting target projects, implementing training data construction projects, building and managing the integrated provision system, and charging fees shall be prescribed by Presidential Decree.

Section 2 AI Technology Development and Industry Revitalization

Article 16 (Support for Introduction and Utilization of AI Technology) (1) The State and local governments may, as necessary, provide support for the following to promote the introduction and utilization of AI technology by enterprises and public institutions:

1. Supporting the development and dissemination of AI technology, AI products, or AI services
2. Providing consulting services for enterprises and public institutions seeking to introduce and utilize AI technology
3. Providing education on AI technology introduction and utilization to employees of SMEs, venture businesses, and micro-enterprises (“SMEs, etc.”)
4. Financial support for SMEs, etc. to introduce and utilize AI technology
5. Other matters prescribed by Presidential Decree to promote the introduction and utilization of AI technology by enterprises and public institutions (2) Matters necessary for support under Paragraph (1) shall be prescribed by Presidential Decree.

Article 17 (Special Support for SMEs, etc.) (1) When implementing various support measures under this Act for AI technology and the AI industry, SMEs, etc. shall be given priority consideration. (2) The government shall strive to encourage SMEs, etc. to participate in the AI industry and shall reflect such considerations in the Basic Plan. (3) The Minister of Science and ICT may support SMEs, etc. in implementing safety and reliability measures under Article 32 and conducting impact assessments under Article 35 to secure AI safety and reliability.

Article 18 (Vitalizing Start-ups) (1) The government may carry out the following projects to encourage start-ups in the AI industry:

1. Identifying, fostering, and supporting AI start-ups
2. Education and training for AI start-up promotion
3. Supporting the commercialization of outstanding AI technology owned by trained professionals (Article 21)
4. Valuing AI technology and providing financial support for start-up funding
5. Providing results of AI-related research and development
6. Fostering institutions and organizations that support start-ups in the AI industry
7. Other projects necessary to vitalize start-ups in the AI industry (2) Local governments may contribute or invest in public institutions or organizations to support start-ups in the AI industry.

Article 19 (Promotion of AI Convergence) (1) The government shall establish and promote measures to facilitate convergence between the AI industry and other industries and to encourage the utilization of AI in all fields. (2) The government may give priority to including R&D tasks related to AI-converged products and services in national R&D projects under the National R&D Innovation Act. (3) The government shall actively support the smooth implementation of temporary permits and regulatory sandboxes for AI-converged products and services under Articles 37 and 38-2 of the Special Act on Promotion of Information and Communications and on Vitalization of Convergence.

Article 20 (Institutional Improvement, etc.) (1) The government shall endeavor to improve related institutions, including legislation, to develop the AI industry and establish a trust-based foundation. (2) To promote institutional improvements under Paragraph (1), the government may provide administrative and financial support for research on laws and systems and the collection of opinions from various sectors of society.

Article 21 (Securing Professional Personnel) (1) The Minister of Science and ICT shall train and support professional personnel related to AI and AI technology according to the measures under Article 23(1) of the Intelligent Informatization Act. (2) The government may implement the following measures to secure overseas professionals in AI and AI technology:

1. Research and analysis of overseas professionals in AI-related universities, research institutes, and companies
2. Building international networks to attract overseas professionals
3. Supporting the employment of overseas professionals in Korea
4. Supporting the overseas expansion of domestic AI research institutes and attracting foreign AI research institutes to Korea
5. Supporting the hosting of international AI organizations and events in Korea
6. Other matters necessary for securing overseas professionals

Article 22 (International Cooperation and Support for Overseas Expansion) (1) The government shall monitor international trends related to AI and promote international cooperation. (2) The government may provide the following support to strengthen the competitiveness of the AI industry and promote its overseas expansion:

1. International exchange of information, technology, and personnel related to the AI industry
2. Collection, analysis, and provision of information on overseas expansion of the AI industry
3. Joint research and development and international standardization of AI technology, products, or services between countries
4. Attraction of foreign capital investment into the AI industry
5. Participation in overseas academic conferences and exhibitions on AI and related promotion and overseas marketing
6. Establishing sales and distribution networks and cooperation systems for exporting AI products or services
7. Monitoring international trends in AI ethics and supporting international cooperation
8. Other matters necessary to strengthen competitiveness and promote overseas expansion of the AI industry (3) The government may, as prescribed by Presidential Decree, entrust or delegate to public institutions or organizations certain tasks under Paragraph (2), and may subsidize the necessary costs.

Article 23 (Designation of AI Clusters, etc.) (1) The State and local governments may promote the functional, physical, and regional clustering of enterprises, institutions, or organizations engaged in research and development of AI and AI technology to enhance competitiveness in AI development and utilization. (2) To facilitate clustering under Paragraph (1), the State and local governments may designate AI clusters as prescribed by Presidential Decree and provide administrative, financial, and technical support. (3) The State and local governments may revoke the designation of an AI cluster in the following cases. However, if the case falls under Item 1, the designation shall be revoked:

1. If the designation was obtained by fraud or other improper means
2. If the designating authority deems it difficult for the AI cluster to achieve its intended purpose (4) To effectively localize clustering under Paragraph (1), the government may establish or designate a dedicated institution to comprehensively support related tasks. (5) The government may contribute or subsidize all or part of the expenses necessary for operating and performing tasks of the dedicated institution under Paragraph (4). (6) Matters necessary for the designation and revocation of AI clusters and for the establishment or designation of dedicated institutions shall be prescribed by Presidential Decree.

Article 24 (Establishing AI Testbeds, etc.) (1) The State and local governments may build and operate facilities, equipment, and installations (“testbeds, etc.”) necessary for testing, evaluating, or verifying (“testing, etc.”) technologies developed or acquired by AI business operators to provide support. (2) The State and local governments may, as prescribed by Presidential Decree, open testbeds, etc. held by certain institutions to AI business operators to facilitate testing, etc. (3) Matters necessary for building, operating, and opening testbeds, etc. shall be prescribed by Presidential Decree.

Article 25 (Promotion of AI Data Centers, etc.) (1) The government shall promote measures necessary to encourage the establishment and operation of data centers used for AI development and utilization (“AI data centers”). (2) The government may perform the following tasks to implement the measures under Paragraph (1):

1. Administrative and financial support for the establishment and operation of AI data centers
2. Support for SMEs, research institutes, etc. to use AI data centers
3. Support for balanced regional development of AI-related infrastructure, including AI data centers

Article 26 (Establishment of the Korea AI Promotion Association) (1) Persons engaging in research and work related to AI, etc. may, with the approval of the Minister of Science and ICT under Presidential Decree, establish or be designated as the Korea AI Promotion Association (“the Association”) to promote AI development and utilization, the AI industry and technology, and to provide education and publicity related to AI. (2) The Association shall be a corporation. (3) The Association shall carry out the following tasks:

1. Promotion and dissemination of AI technology, products, or services
2. Research and statistical surveys on AI, etc.
3. Installation and operation of shared facilities for AI business operators and education for professional personnel training
4. Support overseas expansion for AI business operators and professionals
5. Education and publicity for safe and trustworthy AI development and utilization
6. Tasks assigned by this or other laws to the Association
7. Other projects necessary to achieve the Association’s objectives as stipulated in its articles of association (4) The State and local governments may, within budgetary limits, provide funds or subsidies necessary for the Association’s operations and implementation of its projects. (5) Matters concerning membership eligibility, officers of the Association, and its operations shall be stipulated in the articles of association; matters to be included therein shall be prescribed by Presidential Decree. (6) When the Minister of Science and ICT grants approval under Paragraph (1), it shall be publicly announced. (7) Except as otherwise provided in this Act, the provisions of the Civil Act concerning incorporated associations shall apply mutatis mutandis to the Association.

Chapter 4 AI Ethics and Reliability Assurance

Article 27 (AI Ethics Principles, etc.) (1) The government may enact and publish AI Ethics Principles (“Ethics Principles”) to disseminate AI ethics, including:

1. Safety and reliability to ensure that AI development and utilization do not harm human life, body, and mental health
2. Accessibility ensuring that everyone can freely and conveniently use AI-based products and services
3. Contribution to human life and prosperity in AI development and utilization (2) The Minister of Science and ICT shall collect opinions from various sectors of society to ensure that all those engaged in AI development and utilization can realize these Ethics Principles, establish implementation plans, and make them public, and promote and educate on them. (3) When a central administrative agency or local government formulates or revises AI ethics standards (including laws, standards, guidelines, or any other forms) related to AI, the Minister of Science and ICT may recommend or express opinions on their connection and consistency with the Ethics Principles and the implementation plans under Paragraph (2).

Article 28 (Establishment of Private Autonomous AI Ethics Committees) (1) The following institutions or organizations may establish a Private Autonomous AI Ethics Committee (“Autonomous Committee”) to comply with the Ethics Principles:

1. Educational or research institutions where persons engaged in AI technology research and development belong
2. AI business operators
3. Other AI technology-related institutions prescribed by Presidential Decree (2) The Autonomous Committee may voluntarily perform the following tasks:
4. Checking compliance with the Ethics Principles in AI technology research, development, and utilization
5. Investigating and researching safety, human rights infringements, etc. in AI technology research, development, and utilization
6. Investigating and supervising procedures and outcomes of AI technology research, development, and utilization
7. Providing ethics education related to the Ethics Principles to researchers and employees of the respective institutions or organizations
8. Preparing sector-specific AI ethics guidelines suitable for AI technology research, development, and utilization
9. Other tasks necessary for realizing the Ethics Principles (3) Matters regarding the composition and operation of the Autonomous Committee shall be determined autonomously by each institution or organization. However, it shall not be composed exclusively of one gender, and shall include persons with experience and knowledge in social and ethical matters, as well as persons not affiliated with the institution or organization. (4) The Minister of Science and ICT may prepare and disseminate standard guidelines to ensure the Autonomous Committee’s fair and neutral composition and operation.

Article 29 (Measures for Building Trust in AI) The government shall develop measures to minimize potential risks of AI on citizens’ lives and to build an environment that ensures safe and trustworthy AI use. Such measures may include:

1. Creating a safe and trustworthy AI usage environment
2. Forecasting and predicting social, economic, cultural, and everyday changes caused by AI use, and improving relevant laws and institutions accordingly
3. Supporting the development and dissemination of safety and certification technologies to ensure AI safety and reliability
4. Educating and publicizing safe and trustworthy AI society implementation and AI ethics
5. Supporting the autonomous establishment and implementation of rules by AI business operators for ensuring safety and reliability
6. Supporting and expanding private activities, such as voluntary cooperation, creation of ethics guidelines, etc., by organizations comprising AI business operators and users (“associations, etc.”) to enhance AI safety and reliability
7. Other matters prescribed by Presidential Decree to ensure AI safety and reliability

Article 30 (Support for AI Safety and Reliability Verification and Certification) (1) The Minister of Science and ICT may implement the following projects to support verification and certification (“verification/certification, etc.”) activities voluntarily undertaken by associations, etc. for ensuring AI safety and reliability:

1. Disseminating guidelines for AI development
2. Supporting research on verification/certification, etc.
3. Supporting the construction and operation of equipment and systems necessary for verification/certification, etc.
4. Training professionals needed for verification/certification, etc.
5. Other matters prescribed by Presidential Decree to support verification/certification, etc. (2) The Minister of Science and ICT may, as prescribed by Presidential Decree, provide information or administrative/financial support to SMEs, etc. wishing to obtain verification/certification, etc. (3) AI business operators providing high-impact AI shall endeavor to obtain verification/certification, etc. in advance. (4) When national institutions intend to use high-impact AI, they shall consider products or services based on AI that have obtained verification/certification, etc. as a priority.

Article 31 (Obligations to Ensure AI Transparency) (1) AI business operators providing products or services using high-impact AI or generative AI shall notify users in advance that the product or service is based on such AI. (2) If providing generative AI or products or services using it, AI business operators shall indicate that the results were generated by generative AI. (3) If providing outputs generated by AI systems that are difficult to distinguish from reality—such as voices, images, or videos—AI business operators shall clearly inform users that the results have been created by an AI system. If the output is an artistic or creative work or part thereof, the notification or indication may be made in a way that does not hinder its exhibition or enjoyment. (4) Matters necessary for the notification and indication methods under Paragraphs (1) to (3) and exceptions thereto shall be prescribed by Presidential Decree.

Article 32 (Obligations to Ensure AI Safety) (1) AI business operators whose AI systems exceed a threshold of cumulative computational usage for training as determined by Presidential Decree shall implement the following measures to ensure safety:

1. Identifying, assessing, and mitigating risks throughout the AI lifecycle
2. Establishing a risk management system to monitor and respond to AI-related safety accidents (2) AI business operators shall submit the results of implementing the measures under Paragraph (1) to the Minister of Science and ICT. (3) The Minister of Science and ICT shall determine and announce detailed guidelines for implementing the measures under Paragraph (1) and for submitting the results under Paragraph (2).

Article 33 (Confirmation of High-Impact AI) (1) When AI business operators provide AI or AI-based products/services, they shall review in advance whether the AI constitutes high-impact AI, and if necessary, they may request the Minister of Science and ICT to confirm whether it is high-impact AI. (2) If requested under Paragraph (1), the Minister of Science and ICT shall confirm whether the AI is high-impact AI, and may, if necessary, seek consultation from a specialized committee. (3) The Minister of Science and ICT may develop and distribute guidelines on criteria and examples of high-impact AI. (4) Matters necessary for the confirmation procedures under Paragraph (1) shall be prescribed by Presidential Decree.

Article 34 (Responsibilities of Business Operators Regarding High-Impact AI) (1) AI business operators providing high-impact AI or AI-based products/services shall ensure safety and reliability by implementing measures that include the following, as prescribed by Presidential Decree:

1. Developing and operating a risk management plan
2. Implementing measures to provide, within technically feasible limits, explanations about the final AI output, the main criteria used to derive it, and an overview of the training data used in the AI’s development and utilization
3. Developing and operating user protection measures
4. Ensuring human oversight and supervision of high-impact AI
5. Preparing and retaining documentation confirming the content of measures taken to ensure safety and reliability
6. Other matters deliberated and resolved by the Committee for ensuring the safety and reliability of high-impact AI (2) The Minister of Science and ICT may establish and announce detailed regulations on the measures under Paragraph (1) and recommend AI business operators to comply. (3) If AI business operators implement measures equivalent to those in Paragraph (1) under other laws as prescribed by Presidential Decree, they shall be deemed to have met the requirements of Paragraph (1).

Article 35 (AI Impact Assessment) (1) AI business operators shall endeavor to conduct an assessment (“impact assessment”) in advance of the effect on human fundamental rights when providing products or services using high-impact AI. (2) When national institutions use products or services based on high-impact AI, they shall give priority to those that have undergone impact assessment. (3) Specific matters regarding the content and methods of the impact assessment shall be prescribed by Presidential Decree.

Article 36 (Designation of a Domestic Representative) (1) AI business operators without an address or business office in Korea who meet certain criteria for number of users, sales, etc., as prescribed by Presidential Decree, shall designate a domestic representative in writing and report this to the Minister of Science and ICT. The domestic representative shall:

1. Submit results under Article 32(2)
2. Apply for confirmation on high-impact AI under Article 33(1)
3. Assist in implementing safety and reliability measures under Article 34(1), including verifying the accuracy and timeliness of documentation under Article 34(1)5 (2) The domestic representative shall have an address or business office in Korea. (3) If the domestic representative violates this Act in connection with the matters in Paragraph (1), the AI business operator who appointed that representative is deemed to have committed the violation.

Chapter 5 Supplementary Provisions

Article 37 (Expansion of Financial Resources for Promoting the AI Industry) (1) The State shall prepare measures to continuously and stably secure necessary financial resources to effectively implement the Basic Plan and policies under this Act. (2) The Minister of Science and ICT may encourage public institutions to provide necessary support for AI industry promotion projects. (3) The State and local governments shall take necessary measures to encourage private-sector investment in AI industry promotion. (4) The State and local governments shall strive to allocate and utilize investment resources efficiently, considering the stage of AI industry development.

Article 38 (Surveys, Statistics, and Indicators) (1) The Minister of Science and ICT, in consultation with the Commissioner of Statistics Korea, shall conduct surveys and produce statistics and indicators related to AI, etc., domestically and internationally, linked to the statistics under Article 26-2 of the Framework Act on Science and Technology, to plan, formulate, and implement the Basic Plan and related policies and projects. The produced statistics and indicators shall be published. (2) The Minister of Science and ICT may request data and cooperation from relevant central administrative agencies, local governments, and public institutions in producing the statistics and indicators under Paragraph (1). Unless special circumstances exist, these agencies shall comply. (3) Matters necessary for conducting surveys, producing, managing, and publishing statistics and indicators shall be prescribed by Presidential Decree.

Article 39 (Delegation and Entrustment of Authority and Duties) (1) The Minister of Science and ICT or the heads of relevant central administrative agencies may delegate part of their authority under this Act to heads of their subordinate agencies or to Special Metropolitan City Mayors, Metropolitan City Mayors, Special Self-Governing City Mayors, Do Governors, or Special Self-Governing Province Governors (hereinafter “Mayor/Do Governor”). In such cases, the Mayor/Do Governor may re-delegate the authority to the heads of Si/Gun/Gu (administrative city included under the Jeju Special Self-Governing Province Act). (2) The government may, as prescribed by Presidential Decree, entrust the following tasks to related institutions or organizations:

1. Support projects for AI technology development and utilization under Article 13
2. Selection and support of target projects and implementation of training data construction projects under Article 15(2) and (3)
3. Construction, operation, and management of the integrated provision system under Article 15(4)
4. Matters deemed necessary by the Minister of Science and ICT to promote start-ups under Article 18
5. Support related to verification/certification under Article 30(2)
6. Surveys, statistics, and indicator preparation under Article 38
7. Other tasks prescribed by Presidential Decree to foster the AI industry and spread AI ethics

Article 40 (Fact-Finding Investigations, etc.) (1) If the Minister of Science and ICT discovers or suspects a violation of this Act, or receives a complaint, the Minister may require AI business operators to submit related data or have public officials conduct necessary investigations. (2) For the investigation under Paragraph (1), the Minister of Science and ICT may have public officials enter the offices or workplaces of AI business operators to examine books, documents, or other materials. (3) If the Minister of Science and ICT confirms a violation of this Act following the investigation under Paragraph (1) and (2), the Minister may order the AI business operator to take necessary measures to stop or correct the violation.

Article 41 (Deemed Public Officials for the Purpose of Criminal Penalties) (1) Members of the Committee who are not public officials shall be considered public officials for the purpose of applying Articles 129 to 132 of the Criminal Act (regarding bribery). (2) Executives and employees of public institutions engaged in entrusted tasks under Article 39(2) shall be considered public officials for the purpose of applying Articles 127 and 129 to 132 of the Criminal Act.

Article 42 (Penalties) Any person who, in violation of Article 7(11), discloses confidential information learned in the course of duty or uses it for purposes other than those of duty shall be punished by imprisonment for up to three years or by a fine of up to 30 million won.

Article 43 (Administrative Fines) (1) A fine of up to 30 million won shall be imposed on any person who:

1. Fails to comply with the notification requirement under Article 31(1)
2. Violates Article 36(1) by failing to designate a domestic representative
3. Fails to comply with an order to cease or correct a violation under Article 40(3) (2) Administrative fines under Paragraph (1) shall be imposed and collected by the Minister of Science and ICT as prescribed by Presidential Decree.

Addenda

Article 1 (Enforcement Date) This Act shall enter into force one year after the date of its promulgation. However, the part of Article 2(4)(d) related to digital medical devices shall take effect on January 24, 2026.

Article 2 (Preparatory Actions for Enforcement of This Act) Appointment of Committee members and other necessary measures for enforcing this Act may be taken before its enforcement date.

Article 3 (Transitional Measures for the Dedicated Institution) As of the enforcement of this Act, any institution that was already receiving government funding to operate for the purpose of effectively localizing clusters under Article 23(1), and that meets the requirements prescribed by Presidential Decree regarding organization, personnel, etc., shall be deemed designated as a dedicated institution under this Act.

Reference: https://likms.assembly.go.kr/bill/billDetail.do?billId=PRC_R2V4H1W1T2K5M1O6E4Q9T0V7Q9S0U0