KRSouth KoreaIn forceAI Basic Act

How to comply with the South Korea AI Basic Act

South Korea AI Basic Act

Status
In force
Binding
Yes
Object analysed
AI system
Requirements
7
Next milestone
Jan 2027

In short

Second horizontal law after the EU. High-impact AI, generative AI labelling, local representative for foreign operators, safety of high-compute models.

Steps to compliance

  1. Qualify each AI systemAxes to decide: Category.
  2. Determine your roleDuties vary by role: Developing operator and Using operator.
  3. Apply the 7 requirementsThey focus on: Risk management, Transparency & notice, Synthetic content and Inventory & categorisation.
  4. Prove it with checks11 checks to document, 11 of which also serve AI Act, PL 2338 and CoE Convention.
  5. Track the deadlinesNext milestone: Jan 2027, Announced end of the fine grace period (potential).

Scope and penalties

Kind
Statute
Scope
Operators developing or using AI, including foreign ones above thresholds.
Territorial reach
Activities affecting the Korean market or users.
Penalties
Administrative fine up to KRW 30M; grace period of at least one year on fines.
Jurisdiction
South Korea

Timeline

Dec 26, 2024Passed by the National Assembly
Jan 21, 2025Promulgated
Jan 22, 2026Entry into force
Release
Jan 2027Announced end of the fine grace periodPotential
PastSet in the textPotentialTo verify

Qualifying a system

Classification axes and possible verdicts

Category

High-impact AIGenerative AIOther

Requirements

7 requirements

CodeArticleRequirementApplies toChecks
KR-01Art. 31(1)Notify in advance of high-impact or generative AI use
Developing operatorUsing operatorHigh-impact AIGenerative AI
KR-02Art. 31(2), (3)Label generated content and flag deepfakes
Developing operatorUsing operatorGenerative AI
KR-03Art. 33Self-assess whether the system is high-impact
Developing operatorUsing operator
KR-04Art. 34High-impact measures: risk management, explainability, user protection, oversight, documentation
Developing operatorUsing operatorHigh-impact AI
KR-05Art. 35Fundamental rights impact assessment (best effort)
Using operatorHigh-impact AI
KR-06Art. 32Safety of high-compute models
Developing operator
KR-07Art. 36Domestic representative for foreign operators
Developing operatorUsing operator

Checks to document

Evidence collected for a check counts for every regulation that uses it.

CodeCheckScopeThemesAlso used by
VER-018-D-01Persons informed of the interaction with an AI systemSystem
VER-018-D-03AI-generated content marked as suchSystem
CHK-CATEGORIZATIONAI system tasks and methods are categorized (classifier, generative, recommender)System
VER-003-01Documented and up-to-date risk registerSystem
VER-033-D-02System explanation capability verifiedSystem
VER-008-02System designed to allow human oversight (stop button, override)System
VER-005-01Complete technical documentation compliant with Annex IVSystem
VER-020-D-01FRIA carried out in accordance with Art. 27System
VER-032-G-02Systemic risks assessed and mitigation measures documentedModel
VER-032-G-03Serious incidents documented and reported to the AI OfficeModel
VER-025-F-01Authorised representative appointed by written mandateOrganisation

Themes covered

Frequently asked questions

Who is in scope of AI Basic Act?

Operators developing or using AI, including foreign ones above thresholds. Activities affecting the Korean market or users.

What penalties does AI Basic Act carry?

Administrative fine up to KRW 30M; grace period of at least one year on fines.

When do the AI Basic Act obligations apply?

Jan 21, 2025: Promulgated; Jan 22, 2026: Entry into force; Jan 2027: Announced end of the fine grace period.

Is AI Basic Act binding?

Yes. Kind: statute. Status: in force.

How does AI Basic Act relate to other regulations?

The same checks serve several texts. Shared checks: AI Act (10), PL 2338 (7) and CoE Convention (4).

Related regulations

Official sources

Data checked on Sep 25, 2026. General information, not legal advice. Check the official texts and get advice for your situation.

Run these requirements across all your AI systems

TrustFlow inventories your systems, qualifies them under each regulation and collects evidence once for every referential.

Request a demo