The European Union's delay in releasing guidelines for "high-risk AI" under its AI Act is raising questions about when South Korea will enforce regulations in its own "Framework Act on AI Development and Trust Building" (AI Basic Law).
Some observers suggest that since the EU has officially postponed regulations citing confusion in the IT industry, Korea's implementation schedule may also be adjusted. While the government plans to enforce the AI Basic Law with a grace period of at least one year for certain regulatory provisions, industry voices are calling for additional extensions.
According to the IT industry on the 8th, the European Commission missed its deadline to release high-risk AI guidelines, originally scheduled for February 2nd local time. High-risk AI refers to systems that could significantly affect human health, safety, and fundamental rights, requiring stringent regulation. This corresponds to "high-impact AI" under Korea's AI Basic Law.
The EU guidelines were expected to detail classification systems for high-risk AI and compliance obligations for operators and developers. However, the release appears to have been postponed due to pushback from global big tech companies including OpenAI and Meta, as well as insufficient administrative preparation. EU-focused media outlet Euractiv reported that "the Commission failed to meet the statutory deadline because it needed additional time to review the extensive feedback received from stakeholders and incorporate it into the guidelines."
The delayed guidelines could further postpone full implementation of the AI Act. The Commission had previously proposed in November last year to push back full enforcement of high-risk AI regulations from August this year to December 2027. If discussions on high-risk AI regulations remain unresolved, implementation could be delayed beyond 2028.
Such delays could affect Korea's AI Basic Law, which took effect on January 22nd. Industry observers note that regulatory provisions already include a grace period of at least one year, with possible additional extensions depending on international developments. A senior official at the Ministry of Science and ICT stated, "The EU is a union of multiple countries, making decision-making complex and policy outcomes uncertain." The official added, "Our principle of monitoring AI developments in the EU, the United States, and other countries regarding regulatory enforcement under the AI Basic Law remains unchanged."
Uncertainty persists on the ground despite the regulatory grace period. The AI Basic Law support desk has received 172 consultations in the ten days since opening on January 22nd. Questions about "AI transparency obligations" accounted for the largest share at 56.4%, followed by inquiries about high-impact AI classification at 17%.