The European Union (EU), which had prepared the world’s first artificial intelligence (AI) regulatory legislation, is moving to slow its pace, making it likely that Korea will become the first country in the world to implement an “AI Act” in January next year.

According to the Ministry of Science and ICT on the 14th, Korea will implement the “Framework Act on Fostering Artificial Intelligence and Building Trust” (AI Framework Act) on January 22 next year, after completing the enforcement decree’s legislative notice period by the 22nd of this month. The legislation was drafted by benchmarking the EU’s AI regulatory proposal, but in terms of actual enforcement, Korea will move ahead.

The European Commission on November 19 announced that it would postpone the application of key provisions of the world’s first comprehensive AI regulation, the “AI Act”, and relax personal data protection rules. It also decided to delay from August next year to December 2027 the date by which companies must comply with the EU’s strict rules when using “high-risk” AI that could threaten health, safety, and fundamental rights. The move came as a check on regulation amid concerns that EU member states and companies could fall behind in the AI race, while the United States and China are locked in intense competition for AI supremacy.

While the EU is thus moving to adjust its pace, in Korea the enforcement of the law in January next year is imminent, and there is considerable discontent in the industry. Various AI services are being launched, but companies say more preparation time is needed to respond to the law’s implementation.

In fact, a survey conducted on the 3rd of this month by Startup Alliance on 101 AI startups showed that 98% of the startups had not established a substantive response system related to the AI Framework Act. The most restrictive provisions of the AI Framework Act were cited in the following order: △ reliability and safety certification system (27.7%) △ requirements to ensure dataset transparency (23.8%) △ designation of high-risk AI and obligations for registration and verification (17.8%) △ obligation to label outputs from generative AI (15.8%).

Lee Gi-dae, head of the center at Startup Alliance, said, “The enforcement of the AI Framework Act is imminent, but preparations on the ground are still insufficient.” The Information Technology and Innovation Foundation (ITIF), a U.S. think tank specializing in technology policy, also issued a report in October this year on Korea’s AI Framework Act, pointing out that “an inaccurate assessment of AI risks could halve the policy effectiveness of the industrial promotion component.”

Reflecting such concerns, the Ministry of Science and ICT has stated that during the one-year period in which administrative fines are deferred, it will focus on stabilizing the system rather than strictly applying the law. A ministry official said, “We see next year as a period for the law to take root and plan to focus on guidance so as not to impede AI innovation,” adding, “We will defer administrative fines for at least one year.”

Popular News