Framework Act on the Development of Artificial Intelligence and the Creation of a Foundation for Trust, South Korea

Prepared by Shubhank Suman · Last reviewed 20 September 2026

Document page

CURRENT / AMENDED 2026

Legal character: National AI statute
Region: Asia-Pacific
Original enactment: 21 January 2025
Major amendment: Act No. 21311, 20 January 2026
Current effective version: 21 July 2026

What it covers

South Korea's principal horizontal AI statute covers AI development and industry promotion, governance, ethics, trustworthiness, transparency, safety, high-impact and generative AI, impact assessments, domestic representatives, certification and enforcement. It may apply to conduct outside Korea that affects the domestic market or users.

Why it matters

The amended current version creates operational transparency, safety and high-impact-AI duties, including incident-management and domestic-representative provisions.

Refer to it when

Use it for Korean generative-AI transparency, synthetic-content labelling, high-impact classification, lifecycle safety, impact assessment, domestic representatives and enforcement.

AI, privacy and cybersecurity relevance

AI: Very High
Privacy: Very High
Cyber: Very High

VERY HIGH RELEVANCE

This is South Korea's current national horizontal AI statute.

Privacy

VERY HIGH RELEVANCE

Fundamental rights, transparency, explanation, training-data information, vulnerable groups and impact assessment overlap with privacy and data governance, while Korea's PIPA remains separately applicable.

Cybersecurity

VERY HIGH RELEVANCE

Article 32 creates explicit lifecycle safety, risk-management, technical-control, resilience and incident-management obligations.

Applicability

Relevant AI business operators and covered AI systems, including out-of-country conduct affecting the Korean market or users under the statutory rules.

Current validity

CURRENT: AMENDED 2026 VERSION EFFECTIVE 21 JULY 2026.

Read first

Articles 1–5 → Article 6 → Articles 27–30 → Article 31 → Article 32 → Articles 33–35 → Article 36 → Articles 42–43 → Enforcement Decree.

Provisions / sections covered

Article 2: Definitions

Defines AI, AI systems, generative AI and high-impact uses affecting life, safety, rights, employment, lending, public services, education and transport.

Article 6: AI master plan

Requires a three-year master plan covering policy, training, ethics, investment, trust, fairness, transparency, accountability, safety, accessibility and vulnerable groups.

Article 30: Safety and trustworthiness certification

Supports voluntary verification and certification.

Article 31: AI transparency

Requires relevant notices, identification of generative outputs and recognisable identification of realistic synthetic media.

Article 32: AI safety

For systems above the prescribed compute threshold, requires lifecycle risk identification, assessment and mitigation, incident-management systems and reporting.

Articles 33–35: High-impact AI

Cover classification confirmation, risk management, explanation, criteria and training-data information, trustworthiness measures and impact assessment.

Article 36: Domestic representative

Creates obligations for specified foreign AI business operators.

Articles 42–43: Enforcement

Contain criminal and administrative penalty provisions.

Subject-matter tags describe what the indexed text discusses. They are not legal conclusions.

Document page in the Knowledge Base - full text or PDF where it is available for this record.

This library locates primary and official sources. It does not interpret the law and it is not legal advice.