KR AI Ethics Principles
Republic of Korea Artificial Intelligence Ethics Principles (대한민국 인공지능 윤리원칙)
Government-wide AI ethics principles issued under Article 27 of Korea's AI Framework Act, setting three core values and seven principles applying across the AI lifecycle. Non-binding, addressed to AI suppliers, users, and government.
Jurisdiction
South Korea
Enacted
Aug 21, 2026
Effective
Aug 24, 2026
Enforcement
None (non-binding). Administered by the Ministry of Science and ICT.
Resolved at the 12th Meeting of Ministers Related to Science and Technology on August 21, 2026 and announced jointly by the Ministry of Science and ICT and the Korea Information Society Development Institute (KISDI) on August 24, 2026. Issued under Article 27 of the AI Framework Act as a government-wide, society-wide self-regulatory norm (자율규범) with no legal binding force; it does not displace obligations imposed by statute.
Korea Policy Briefing (Government of the Republic of Korea)Why It Matters
Supplies the reference vocabulary Korean ministries are expected to use when interpreting the open-textured duties of the AI Framework Act, without itself creating obligations. Relevant as interpretive context for operators already within the Framework Act's scope rather than as a compliance requirement in its own right.
Recent Developments
Adopted August 21, 2026 at the 12th Meeting of Ministers Related to Science and Technology and announced August 24, 2026, following a public debate process convened by the Ministry of Science and ICT to gather views from across sectors. Each principle is accompanied by common standards and actor-specific roles for suppliers, users, and government across the AI lifecycle.
At a Glance
Applies to
Who Must Comply
- AI suppliers (developers and providers) operating in Korea
- Users of AI systems
- Government bodies and society at large
Safety Provisions
- Safety principle applying across the full AI lifecycle, with roles allocated separately to suppliers, users, and government
- Human-centeredness principle grounded in the core value of human dignity
- Trustworthiness and transparency principles covering disclosure about AI systems and their operation
- Accountability principle allocating responsibility among actors across the AI lifecycle
- Privacy protection and fairness/inclusivity principles
Compliance & Enforcement
Penalties
None (non-binding self-regulatory norm)
View on map
South Korea
Focus Areas
Cite This
APA
South Korea. (2026). Republic of Korea Artificial Intelligence Ethics Principles (대한민국 인공지능 윤리원칙).
Related Regulations
Korea AI Act
First comprehensive AI legislation in Asia-Pacific and second in the world after EU. Regulates "High-Impact AI" in healthcare, energy, nuclear, transport, government, and education sectors. Requires transparency notifications, content labeling for generative AI, and fundamental rights impact assessments. Notable for lower penalties than EU AI Act and absence of prohibited AI practices.
South Korea Deepfake Law
South Korea's world-strictest deepfake law: 7 years for creating/distributing, 3 years for possessing/viewing deepfake sexual content. Even viewing is criminal.
China CSL Amendments
First major revision of China's foundational Cybersecurity Law since 2017. Introduces formal AI governance provisions, significantly increases penalties, and expands extraterritorial application to all cybersecurity violations.
AU National AI Plan
National AI policy roadmap replacing previously proposed mandatory AI guardrails. Focuses on leveraging existing legal frameworks rather than new mandatory requirements. Establishes the Australian AI Safety Institute (AISI) to monitor, test, and share information on AI risks and harms.
Brunei PDPO
Brunei's personal data protection order requiring DPIA and imposing penalties up to 10% Brunei turnover or $1M.
India DPDP Act
STRICTEST children's provisions in APAC. Children = under 18; verifiable parental consent MANDATORY; PROHIBITION on tracking, behavioral monitoring, targeted advertising to children.
Last updated August 30, 2026. Verify against primary sources before relying on this information.