South Korea's AI Advertising Law in 2026: Labeling Rules, KFTC Enforcement and Misleading-Claim Penalties

South Korea's 2026 AI Advertising Turn

South Korea is moving to make AI disclosure a condition of advertising. Through 2026 the government plans to require that AI-generated or AI-assisted advertising be clearly labeled, and to tighten how false, exaggerated and AI-fabricated claims are treated under existing fair-advertising law. For any brand running campaigns in the Korean market — directly or through a local agency — the change reframes a creative decision as a compliance decision: if AI was used to generate or materially alter an ad, that fact is expected to be disclosed.

The shift matters because it sits on top of one of Asia's more active advertising-enforcement regimes. The Korea Fair Trade Commission (KFTC) already polices false and misleading advertising under the Act on Fair Labeling and Advertising, and the 2026 measures extend that posture to synthetic media, virtual endorsers and AI-fabricated expert recommendations. The result is a regime where both how an ad was made and what it claims are in scope.

"AI disclosure turns a production detail into a labeling obligation. In Korea's framing, an undisclosed AI endorser or AI-generated visual is not a stylistic choice — it is potentially unfair advertising.
— AuditSocials analysis of South Korea's 2026 AI advertising measures"

This guide explains the AI labeling rule, the laws behind it, how misleading-claim and virtual-endorser rules apply, what platforms must do, how the KFTC enforces, and the concrete steps advertisers should take. For the jurisdiction baseline, see our South Korea advertising regulations guide, and define terms in the compliance glossary.

The AI-Generated Content Labeling Rule

The core of the 2026 change is a disclosure duty: AI-generated or AI-assisted advertising content is expected to be labeled so that a viewer can tell the material was made with AI. The obligation is framed broadly — it reaches the people who produce, edit or upload the content, not only the brand that commissions it.

Who Must Label, and What

For advertisers, the practical implication is that AI provenance has to be tracked through the production chain. If a creative is generated with a text-to-image or text-to-video model, or a real asset is materially altered with AI, the safe assumption is that it needs a visible AI indicator when used in a Korean campaign. Because the precise format and threshold of the label are being set through guidelines rather than a single bright-line rule, document how each asset was produced and default to disclosure where AI played a generative role. Audit how your creative reads against AI-disclosure expectations with the AI Compliance Audit.

The Laws Behind the Rules

The 2026 measures are not a single statute but a layered framework. Understanding which instrument does what helps advertisers anticipate where obligations harden into enforceable duties.

The Legal Stack

This layered design means the disclosure duty is reinforced from more than one direction: a transparency obligation under the AI Basic Act, a labeling and platform-duty layer under the network and broadcasting framework, and an unfair-advertising backstop under the Act on Fair Labeling and Advertising enforced by the KFTC. Advertisers used to single-regulator regimes should map which body owns which obligation. The structure is comparable in ambition to the EU approach to synthetic-media transparency under the AI Act — see our EU DSA and digital-rules guide for the European parallel. Track amendments as they are finalized on the Policy Change Tracker.

Misleading Claims and Virtual Endorsers

The 2026 measures are not only about a visible AI tag. They extend Korea's long-standing prohibition on false and misleading advertising to AI-fabricated endorsements and exaggerated claims dressed up as expert opinion.

Where AI Meets Misleading Advertising

• Undisclosed virtual endorsers: The KFTC and the Ministry of Food and Drug Safety are clarifying standards so that advertising in which an AI-generated virtual human or expert recommends a product, without disclosing that the endorser is virtual, can be treated as unfair labeling or advertising.
• Fake expert recommendations: The government plans punitive damages of up to five times the losses for knowingly distributing false AI-generated ads, with fabricated expert endorsements — for example AI-generated "doctor" recommendations — a particular target.
• Exaggerated and false claims: The base prohibition under the Act on Fair Labeling and Advertising still applies: claims must not deceive or mislead consumers, and material facts cannot be concealed, whether or not AI was used to make the ad.
• Health and medical context: Product categories overseen by the MFDS — health functional foods, cosmetics and medical-adjacent products — face heightened scrutiny when AI personas or synthetic "experts" are used to lend false authority.

The throughline is authority. Korea's concern is that AI lets advertisers manufacture the appearance of a credible human endorser or expert, and the rules are designed so that synthetic authority must be disclosed and substantiated. For advertisers using creators or endorsers, this overlaps with disclosure duties familiar from other markets; check endorsement language with the Disclosure Checker and pressure-test claims with the Keyword Risk Checker before a campaign goes live.

What Platforms Must Do

South Korea's approach does not place the duty on advertisers alone. Platform operators are expected to make compliance possible and to police it, which changes how labeling appears in practice.

Platform Obligations

For advertisers, the platform layer is good news and a constraint at once. It is good news because the major platforms are building standardized AI-disclosure tools — Meta, TikTok, YouTube and Google already operate AI-content labeling systems globally — so the Korean obligation can often be met using native controls rather than bespoke design. It is a constraint because platform-applied labels may be automatic and hard to remove, so an advertiser cannot quietly strip an AI tag to make a creative look more authentic. Ground each platform's own rules with our Meta ad policies and TikTok guidelines, and compare approaches in the platform comparison.

KFTC Enforcement and Penalties

The teeth behind the 2026 measures come from the KFTC's existing enforcement toolkit under the Act on Fair Labeling and Advertising, layered with new AI-specific duties. Advertisers should understand the range of consequences rather than fixate on a single fine figure.

The Enforcement Toolkit

• Corrective orders: The KFTC can order an advertiser to stop a violating advertisement and, in some cases, to publish a corrective notice.
• Administrative surcharges: Monetary surcharges are calculated on relevant revenue under standards set by Presidential Decree; the KFTC amended its administrative-fine notification with changes effective January 3, 2025, refining how relevant revenue and fixed-amount fines are assessed.
• Criminal referral: Serious false or deceptive advertising can be referred for criminal liability, not only administrative penalties.
• Punitive damages: The government plans punitive damages of up to five times the losses for knowingly distributing false AI-generated ads, with fabricated expert endorsements and deceptive health claims a particular target.

Because the surcharge is tied to relevant revenue rather than a flat cap, exposure scales with the size of the campaign and the product line behind it, which makes large, sustained campaigns the highest-risk profile. The prudent reading for 2026 is that AI disclosure and claim substantiation are not box-ticking items but the difference between a compliant campaign and a corrective order plus surcharge. Keep a defensible record of how each asset was produced and how each claim is substantiated, and monitor enforcement trends on the Policy Change Tracker.

What Advertisers Should Do

The practical response is process, not panic. The 2026 measures reward advertisers who can show how an asset was made and why a claim is true.

Compliance Checklist

• [ ] Inventory which creatives are AI-generated or materially AI-altered, and flag them for disclosure in Korean campaigns
• [ ] Default to a visible AI indicator where AI played a generative role; do not remove platform-applied AI labels
• [ ] Disclose when an endorser, presenter or "expert" is an AI-generated virtual human
• [ ] Substantiate every health, performance or comparative claim with evidence on file
• [ ] Apply heightened care for MFDS-regulated categories (health functional foods, cosmetics, medical-adjacent products)
• [ ] Use each platform's native AI-disclosure controls to meet the labeling duty consistently
• [ ] Keep a production and substantiation record for each asset in case of a KFTC inquiry
• [ ] Track amendments to the Act on Fair Labeling and Advertising and the AI Basic Act as they finalize

Run AI-generated creative through the AI Compliance Audit, screen claim language with the Keyword Risk Checker, and for multi-market campaigns use the Legal Compliance Scan. For sector-specific exposure, see our healthcare social media compliance guide.