April 28, 2026: Finalization of EU AI Act technical standards. Intersecting with the 100th day of the enforcement of South Korea's Artificial Intelligence Basic Act, the global AI regulatory landscape is entering a new phase. [Image = AI-generated image]
The 'Brussels Effect' of Regulation Lands in Seoul
In April 2026, the global AI regulatory landscape entered a historical phase where two major axes collide.
On April 28, 2026, news that the European Union (EU) finalized the technical standards and governance framework for the 'AI Act'—the world's first comprehensive artificial intelligence regulation—sent deep shockwaves through the AI industry. Concurrently, South Korea has been enforcing its own 'Framework Act on the Promotion of Artificial Intelligence and Creation of Trust Bases' (hereinafter the AI Basic Act) since January 22. South Korea is evaluated as highly likely to become the earliest case among nations to fully and practically apply AI-related norms.
With both regulatory frameworks beginning to operate simultaneously, South Korean AI companies face an unprecedented burden of dual compliance. This report comprehensively analyzes the significance of the finalization of the EU AI Act technical standards, the current status of South Korea's AI Basic Act at 100 days, the points of collision between the two regulatory frameworks, and the response tasks for businesses and the government.
▌ 1. Finalization of EU AI Act Technical Standards: Materialization of the 'Brussels Effect'
Current Position of Phased Enforcement
The EU AI Act was not enacted all at once. Beginning with its entry into force in August 2024, it became the world's first comprehensive AI regulation, with prohibited AI practices becoming illegal on February 2, 2025, and general-purpose AI (GPAI) model obligations taking effect on August 2, 2025.
The finalization of the technical standards and governance framework on April 28 marks a core milestone in this phased implementation structure. Regulations on prohibited AI use cases under the AI Act are scheduled to apply first starting in August 2026, followed by regulations on general-purpose AI models in 2027, and full regulations on high-risk AI systems in 2028. With the technical standards now finalized, the specific compliance pathways for regulations slated for application this August have finally become clear.
Variables of the 'Digital Omnibus': Postponement or Mitigation?
However, the trajectory of the EU AI Act is not linear. A 'Digital Omnibus' proposal to delay compliance deadlines for certain high-risk AI systems is under discussion, making April and May a period to watch for potential schedule postponements. The core contents of this proposal are already taking shape: the enforcement of high-risk AI regulations in sensitive fields such as employment and law enforcement is delayed by up to 16 months from August 2026 to December 2027, and high-risk AI regulations embedded in medical devices are delayed by up to 12 months to August 2027.
The Digital Omnibus goes beyond mere delays. It expands mitigation measures previously applied only to small and medium-sized enterprises (SMEs) to mid-sized companies with fewer than 750 employees, bringing an additional approximately 8,250 companies under simplified benefits, and includes exemptions from EU database registration obligations for AI used for limited internal corporate purposes.
Intra-EU conflicts are also drawing attention. While France and Germany argue for regulatory easing to protect their domestic AI industries, Nordic countries demand the application of stricter ethical standards. How the Digital Omnibus concludes amidst these fractures serves as a direct variable for response strategies of companies worldwide, including those in South Korea.
The Reality of the 'Brussels Effect': Extraterritorial Reach
Running through the core of all these discussions is the so-called 'Brussels Effect.' The formation of the EU's AI norms and governance standardizes strict regulations across the entire EU market, requiring countries wishing to enter the EU market to comply as well. This is evaluated as having a 'Brussels Effect,' meaning that regulations formed by the EU exert binding power globally.
The level of sanctions—imposing fines for high-risk AI system violations of up to 35 million euros or 7% of total worldwide annual turnover, whichever is higher—acts as powerful pressure making this regulation an 'obligation' rather than an 'option'.
▌ 2. South Korea's AI Basic Act at 100 Days: Report Card of the World's First Full Enforcement
Historical Significance and Structure of Enforcement
On January 22, 2026, South Korea became the world's first nation to fully enforce AI regulations. Considering that the EU has staged the application timing of its high-risk AI regulations, South Korea is expected to be the country that substantially applies AI regulations in full first.
The core design principle of the law is the balance between promotion and regulation. According to an analysis by the Korea Legislation Research Institute, the AI Basic Act was designed with the direction of