The 'Basic Act on the Development of Artificial Intelligence and Creation of a Trust Base (Basic Act on Artificial Intelligence)' passed the National Assembly with minimal regulations and the exclusion of fine provisions, and it is expected that regulatory uncertainty in artificial intelligence (AI) businesses will be resolved and related business promotion will gain momentum.

The Artificial Intelligence Basic Act (AI Basic Act) passed the National Assembly plenary session on December 26.

The Basic AI Act is a bill that has been discussed for over four years, collecting various opinions, since it was first proposed in the National Assembly in July 2020. The 22nd National Assembly merged 19 bills through agreement between the ruling and opposition parties, and passed them in the plenary session after going through the Defense Committee (November 26, 2024) and the Legislation and Judiciary Committee (December 17, 2024).

The AI Basic Law is the world's second basic law enacted after the European Union in June 2024, and the United States enacted an artificial intelligence executive order in October 2023.

The 'Basic AI Act' is scheduled to take effect in January 2026 after a one-year grace period following deliberation and promulgation by the State Council. The government also plans to take follow-up measures, such as establishing lower-level laws and guidelines, early (in the first half of 2025) so that the law can quickly settle into the market.

■ Establishing a promotion system for national AI development and creation of a trust foundation

Looking at the main contents, Article 2, Paragraph 2 defines an 'artificial intelligence system' as an artificial intelligence-based system that infers results such as predictions, recommendations, and decisions that affect real and virtual environments for given goals with various levels of autonomy and adaptability.

The Minister of Science and ICT may establish and implement a basic plan for artificial intelligence every three years (Article 6) to strengthen the nation's artificial intelligence competitiveness by collecting opinions from relevant ministries and local governments.

In addition, the legal basis for the operation of the National Artificial Intelligence Commission (Chairman: President) launched last September was established (Article 7), and the basis for the operation of the Artificial Intelligence Safety Research Institute as a specialized agency to protect the lives, bodies, and property of citizens from risks that may arise in relation to artificial intelligence was established (Article 12).

■ Support for fostering the artificial intelligence industry, including artificial intelligence research and development, learning data, artificial intelligence data centers, and integrated complexes

In order to foster the artificial intelligence industry, the basis for government support was established for research and development support (Article 13), standardization (Article 14), establishment of learning data policies (Article 15), and support for the introduction and use of artificial intelligence (Article 16).

In addition, we sought to support the innovative development of the artificial intelligence ecosystem in Korea through designation of an artificial intelligence complex (Article 23), promotion of artificial intelligence data center policies (Article 25), and promotion of artificial intelligence convergence (Article 19), and we sought to secure human resources who will lead the artificial intelligence era through securing artificial intelligence experts (Article 21), special support for small and medium-sized enterprises (Article 17), and revitalization of startups (Article 18). It also stipulates provisions to promote the growth of small and medium-sized enterprises and start-ups.

■ Creating a safety and trust foundation for high-impact artificial intelligence and generative artificial intelligence

In order to prevent problems that may arise in advance due to the technical limitations of artificial intelligence, misuse, etc., high-impact artificial intelligence and generative artificial intelligence were defined as subjects of regulation, and the obligation to ensure transparency (Article 31), the obligation to ensure safety (Article 32), and the responsibilities of business operators (Article 34) were stipulated.

In addition, the government has also established a basis for support for private-sector autonomous AI safety and reliability verification and certification (Article 30) and AI impact assessment (Article 35).

■ High-impact artificial intelligence

The Basic AI Act introduced the term “high-impact artificial intelligence” instead of “high-risk artificial intelligence.” According to Article 2, Paragraph 4, “high-impact artificial intelligence” means an artificial intelligence system that has a significant impact on or is likely to pose a risk to human life, physical safety, and basic rights, and is utilized in any of the following areas.

- Supply of energy according to Article 2, Paragraph 1 of the ‘Energy Act’
- Production process of drinking water according to Article 3, Paragraph 1 of the Drinking Water Management Act
- Establishment and operation of a system for provision and use of health care services pursuant to Article 3, Paragraph 1 of the ‘Basic Health Care Act’
- Development and use of medical devices according to Article 2, Paragraph 1 of the ‘Medical Devices Act’ and digital medical devices according to Article 2, Paragraph 2 of the ‘Digital Medical Products Act’
- According to Article 2, Paragraph 1, Subparagraph 1 of the ‘Atomic Energy Facilities Protection and Radiation Disaster Prevention Act’Safe management and operation of nuclear facilities under Article 2 of the same Act, including other nuclear materials
- Analysis and use of personal information related to physical, physiological, and behavioral characteristics that can identify an individual, such as face, fingerprints, irises, and palm veins, for criminal investigations or arrests
- Judgments or evaluations that have a significant impact on an individual's rights and obligations, such as hiring or loan screening
- Major operations and management of transportation means, transportation facilities, and transportation systems according to Article 2, Paragraphs 1 through 3 of the 'Traffic Safety Act'
- Decision-making by the state, local governments, public institutions under Article 4 of the Act on the Operation of Public Institutions, etc. that affect the public, such as verification and decision of qualifications necessary for providing public services or collection of costs
- Student evaluation in early childhood education, elementary education, and secondary education according to Article 9, Paragraph 1 of the ‘Basic Education Act’
- Other areas that have a significant impact on the safety of human life and body and the protection of basic rights, as determined by Presidential Decree.

In relation to high-impact artificial intelligence, AI business operators must review in advance whether the AI in question is high-impact artificial intelligence when providing AI or products/services utilizing AI, and, if necessary, may request the Minister of Science and ICT to confirm whether it is high-impact artificial intelligence.

In order to secure the safety and reliability of high-impact artificial intelligence, AI business operators shall establish and operate risk management plans in accordance with Article 34, Paragraphs 1 through 3; establish and implement explanation plans for the final results derived from artificial intelligence within the scope technically possible, key criteria used to derive the final results from artificial intelligence, and an overview of learning data used in the development and use of artificial intelligence; establish and operate user protection plans; and provide information on high-impact artificial intelligence.△Prepare and store documents that can confirm the contents of measures to ensure safety and reliability; △Implement other measures such as matters deliberated and decided by the committee to ensure the safety and reliability of high-impact artificial intelligence.

■ Generative Artificial Intelligence

Article 2, Paragraph 5 defines generative artificial intelligence as an artificial intelligence system that creates text, sound, images, videos, and other various outputs by imitating the structure and characteristics of input data.

The Basic AI Act requires that when an AI business operator intends to provide a product or service using generative artificial intelligence, it must notify users in advance that the product or service is operated based on generative artificial intelligence.

In addition, AI service providers must indicate that the results of their products or services were generated by generative artificial intelligence, and in particular, when providing virtual sounds, images, or videos that are difficult to distinguish from reality, they must notify or indicate in a way that allows users to clearly recognize that the results were generated by an artificial intelligence system.

■ In case of violation, a fine of up to 30 million won

If the Minister of Science and ICT discovers or suspects a violation of the following provisions of Article 40, Paragraphs 1 through 3: △obligation to display generative artificial intelligence results; △obligation to notify and display virtual results that are difficult to distinguish from reality; △obligation to take safety measures and submit results for artificial intelligence systems whose cumulative computational volume used for learning exceeds the standard prescribed by Presidential Decree; △obligation to take safety and reliability measures for high-impact artificial intelligence: The Minister may conduct a fact-finding investigation and, if a violation is acknowledged, may order the cessation or correction of the violation.

In addition, in accordance with Article 43, if △the suspension order or corrective order is not complied with, △the obligation to provide prior notice regarding high-impact artificial intelligence or generative artificial intelligence is not fulfilled, △a domestic agent is not designated, a fine of up to 30 million won may be imposed.

■ Also applies to acts committed overseas

The Framework Act on AI also applies to acts committed overseas that affect the domestic market or users, and the Framework Act on AI is excluded from application to artificial intelligence that is developed and used solely for national defense or national security purposes, as determined by Presidential Decree.

A legal community official said, “The Basic Act on Artificial Intelligence appears to be excluding criminal punishment and fine provisions for AI business operators and establishing the minimum necessary regulations,” adding, “As the overall regulatory framework becomes clearer, the legal uncertainty that AI business operators have faced in the process of promoting their businesses will likely be alleviated, and the predictability of AI-related regulations will likely increase.”