Artificial intelligence (AI) is evolving into an essential technology in everyday life and the business environment. The highly autonomous AI technology is solving complex problems and simplifying everyday life, and is expected to be actively used in tackling major challenges facing humanity, including social, scientific, medical, and food problems. Currently, AI is a major driver of national development, and the competition to dominate AI technology is intensifying not just between companies but also on a global level. Korea is the world’s sixth most competitive nation in terms of AI capacity (Tortoise Intelligence, The Global AI Index, 2023) and is highly rated in the categories of infrastructure, development, government strategy, and scale.

In 2025, the size of the Korean AI market is estimated to grow by 12.1% from the previous year and reach KRW 3.43 trillion. The market is expected to grow at an annual average growth rate of 14.3% to achieve KRW 4.46 trillion in 2027 (IDC Semiannual Artificial Intelligence Tracker, 2023). Most notably, the country’s AI application industry shows promising outlook in major industries such as information and communication, manufacturing, medical, public/defense, finance, and education, and the industry is expected to attract more investment aimed at strengthening corporate competitiveness and creating new business opportunities.

The level of Korea's AI-related technology has grown from 81.6% of the global leader in 2018 to 88.9% in 2022, and the country’s technology has advanced faster than any other major country in the five years (ICT Technology Level Survey and Technology Competitiveness Analysis Report, Institute for Information & communication Technology Planning & uation (IITP), 2023). In terms of overall technology level, the country's level of AI technology has improved by 7.3 percentage points from 81.6% in 2018 to 88.9% in 2022, compared to the United States (100%), the global leader in AI technology. More specifically, the pace of development of AI technology in application stage is more than twice as fast as that of major countries. In addition, Korea's AI technology gap with the US in each field is 1.3 years in the category of learning intelligence (reduced by 0.7 years from 2018), 1.5 years in the category of single intelligence (reduced by 0.5 years from 2018), and 1.0 year in the category of complex intelligence (reduced by 1.0 year from 2018). The results show that Korea closed the technology gap faster than any other country during the period.

Visual intelligence accounted for the highest share of AI technologies held by Korean companies (30.1%), followed by reasoning and knowledge representation (25.1%) and language intelligence (12.8%). These technologies contributing greatly to automating operations and making them more efficiency, driven by AI-based image recognition and natural language processing technologies in various fields. The improvement of AI's logical thinking and reasoning capabilities, and the advancement of related infrastructure and data processing capabilities are expected to accelerate innovation in a wider range of industries.

In December last year, the Korean National Assembly passed the “Framework Act on the Advancement of Artificial Intelligence and the Establishment of a Foundation for Trust” (hereinafter referred to as the “Framework Act on AI”) at the plenary session. The passage made Korea the world’s second state, after the European Union (EU), to enact a framework act on AI. The Framework Act on AI is positively reviewed as having established a government-supported governance system for promoting AI technology and the industry and laid the foundation for enhancing the country’s AI competitiveness by creating a responsible AI ecosystem. However, there are concerns that the details of regulations and the ambiguity of obligations will add uncertainties to the business environment. In light of the situation, this article aims to review the differences with the European Union’s AI Act, the concerns that arise from these differences, and the outlook for the future.

Korea’s Framework Act on AI and the EU AI Act are common in the sense that they both take a risk-based approach and aim to preempt potential risks. However, the Framework Act on AI differs in terms of the targets and scope of regulation and the severity of sanctions. First, unlike the EU AI Act, which defines four levels of risk, including unacceptable risk, high risk, limited risk and minimal risk, the Framework Act on AI focuses on the “impact” of AI and distinguishes between “high-impact AI (AI systems harboring the risk of seriously affecting or endangering human lives, physical safety and basic rights) and “generative AI systems (that imitate the structure and characteristics of input data to generate various outputs such as text, sound, pictures, and videos).” The Framework Act on AI includes both AI developers (parties that develop and provide AI) and AI users (parties that provide products or services using AI) as AI businesses that are subject to regulation and impose common obligations. This is different from the EU AI Act, which distinguishes between providers, deployers, importers, and distributors of high-risk AI and applies different obligations.

Second, the Framework Act on AI grants exceptions for AI developed and used only for the purpose of national defense and security, which is narrower than the EU AI Act allowing several exceptions, such as AI for national defense and security, pre-launch research, test and development, scientific research and development, and purely private and non-professional activities.

Third, the Framework Act on AI only stipulates a fine of up to KRW 30 million, regardless of the scale of the company, technology, or service, as opposed to the EU AI Act, which imposes fines of up to 7% of global sales or EUR 35 million, and the severity of the sanctions imposed depending on the company's size.

In summary, the main issues raised in the Framework Act on AI, when compared with the EU AI Act, include ambiguity in the content and targets of regulations, narrow exceptions, and lack of strictness and differential application standards for preemptive sanctions. There are also concerns that these uncertainties could lead to adverse effects. However, the Framework Act on AI is significant in that it provides a basis for establishing policies to promote AI technology and industry at the government level. The scope and standards for the implementation of obligations and others will be established in subordinate statutes, and the contents of risk management, ensuring the safety and reliability of high-impact AI, and user protection will be specified in guidelines and notices. A public-private joint consultative body (subordinate statute preparation group) consisting of industry insiders, scholars, and legal experts was launched by the Ministry of Science and ICT ahead of the Framework Act on AI’s enforcement in January 2026, with the aim to prepare a draft of the subordinate statute by June this year. The government and related agencies are expected make swift progress to resolve uncertainties and help companies predict policy directions, thereby promoting large-scale public and private investments.