Meta, Google, and TikTok each require some form of AI disclosure in ads, but the trigger and the label differ. Meta's rule covers synthetic and AI-edited creative, Google's covers election ads and, since July 2026, offers a general labeling control, and TikTok requires disclosure of realistic AI content across organic posts, branded content, and paid ads.
What does Meta require for AI-generated ads?
Meta requires an "AI Info" disclosure label on ads containing AI-generated or significantly AI-edited content. Starting June 1, 2026, Meta added automated detection for social-issue, election, or political ads. For those categories, advertisers must disclose when a photorealistic image, video, or realistic audio was created or edited with third-party generative AI tools to depict a real person doing something they did not do, or a realistic-looking person or event that does not exist. Once Meta's systems detect qualifying content, the AI Info label appears in the ad's "About this Ad" panel with no advertiser action required. Meta also reads embedded Content Credentials from the Coalition for Content Provenance and Authenticity (C2PA), which major generative tools write into their output files, and applies the label automatically when it finds them.
What does Google require for AI-generated ads?
Google requires disclosure of synthetic or digitally altered content in election ads, a rule in place since 2023. On July 9, 2026, Google extended AI disclosure to ordinary commercial ads. Advertisers using Google's own generative AI ad tools get the disclosure applied automatically to the ad's My Ad Center panel, and advertisers using outside AI tools have a control, across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center, and Ads Editor, to mark an individual image or video asset as created or edited with AI. Google describes the rollout as global but gradual, and whether a label shows on the ad itself, not just in My Ad Center, depends on local law.
What does TikTok require for AI-generated ads?
TikTok requires an "AI-generated" label on any content, organic or paid, that a creator or advertiser identifies as fully AI-generated or significantly AI-edited, under a policy TikTok's Newsroom describes as intended to help viewers contextualize realistic synthetic media. For ads, advertisers also apply a mandatory ad-level "AI Disclosure" tag inside TikTok Ads Manager, and the platform then shows a visible "Contains AI-generated content" label on the ad. TikTok integrated C2PA Content Credentials in 2025 and can auto-detect and auto-label undisclosed AI content, and an auto-applied label cannot be removed by the advertiser once live. The branded-content toggle is a separate mechanism, so a sponsored post with AI-generated visuals can need both the paid-partnership disclosure and the AI-generated label.
What does the FTC require for AI-generated content?
The FTC has not issued a comprehensive AI disclosure rule. It has said that existing advertising, endorsement, deception, and testimonial standards apply to AI-generated commercial content. The 2023 update to the Endorsement Guides set the standard for disclosure at "unavoidable," as covered in the FTC disclosure rules 2026 guide.
The FTC's separate rule on fake and AI-generated consumer reviews sits beside that framework. The final rule, effective October 21, 2024, bans creating, buying, or distributing reviews or testimonials that misrepresent the identity or experience of the reviewer, including AI-generated reviews attributed to a person who does not exist. An AI-generated spokesperson or testimonial therefore has to clear two bars: it must not pass as a real person's genuine experience, and any material connection to the advertiser must be disclosed. Platform AI labels do not satisfy the federal requirement on their own. The agency's endorsement guidance is the controlling reference, and enforcement runs through Section 5 of the FTC Act, as laid out in Operation AI Comply.
How do elections and deepfakes change the risk?
Election-related AI content is the fastest-growing enforcement category and carries the most cross-jurisdiction exposure. A brand that does not run political ads can still be affected through public-figure references and influencer activity during election cycles.
| Bucket | What it covers | Where scrutiny comes from |
|---|
| Synthetic political advertising | AI-altered political ads, especially close to an election | State laws; California, Texas, Michigan, and Minnesota have passed laws, and some face court challenges |
| Public figure manipulation | Synthetic likeness, voice, or video of candidates and officials | X, Meta, TikTok, and YouTube apply their strictest protocols |
| Election integrity content | Synthetic content that misstates voting procedures, polling places, or results | Platforms, the FTC, state attorneys general, and the Department of Justice |
Set a higher review threshold for any AI content that touches these areas.
What do New York and EU laws add for AI ads?
New York requires a conspicuous disclosure in any ad that includes a synthetic performer, and the EU AI Act requires machine-readable marking of AI-generated content and disclosure of deepfakes. Both apply based on who the ad reaches, not where the advertiser is based.
What does New York's synthetic performer law require?
New York General Business Law Section 396-b requires a conspicuous disclosure in any advertisement that includes a synthetic performer, defined as a digitally created human likeness including one generated with AI, whenever the ad reaches a New York audience. Governor Kathy Hochul signed the law, S.8420-A/A.8887-B, on December 11, 2025, and it took effect June 9, 2026. Civil penalties are $1,000 for a first violation and $5,000 for each subsequent violation. The trigger is the presence of a synthetic performer, not whether the ad is otherwise misleading, and New York is the first state with a law of this kind aimed specifically at advertising.
What does the EU AI Act require for AI-generated ads?
Article 50 of the EU AI Act requires that AI-generated or manipulated content be marked in a machine-readable format so it can be detected as artificial, and requires deepfake content shown to EU audiences to be disclosed. These transparency obligations, confirmed in the European Commission's guidance, became applicable on August 2, 2026, with a limited grace period to December 2, 2026 for the marking and detection duty on systems already on the market before August 2026. The EU's AI Omnibus delayed high-risk system rules to December 2, 2027, but the transparency obligations were not part of that delay. A deepfake under Article 50 is content that resembles an existing person, object, place, or event, closely enough to be believed authentic, and was generated or manipulated by AI. Non-compliance carries fines of up to €15 million or 3% of worldwide annual turnover, whichever is higher, and the duty applies to any organization whose AI-generated content reaches EU users. The European Commission's AI Act page is the official starting point, and a multi-market brand also needs a state-by-state review.
How do the five ad disclosure regimes compare?
| Regime | What triggers disclosure | Where the disclosure appears | Penalty for non-compliance |
|---|
| Meta | AI-generated or significantly AI-edited creative (automated detection for political ads since June 2026) | "AI Info" label in the ad's About panel | Ad disapproval or removal |
| Google | Synthetic or altered content in election ads (mandatory); any AI-generated asset (advertiser-controlled tag, rolling out from July 2026) | My Ad Center panel; on-ad label where local law requires | Ad disapproval |
| TikTok | Realistic AI-generated or AI-edited content, organic or paid | "AI-generated" label; separate ad-level "AI Disclosure" tag | Content removal, ad rejection, account review |
| New York GBL 396-b | A synthetic performer appears in an ad reaching New York | Conspicuous disclosure within the ad | $1,000 first violation, $5,000 per subsequent violation |
| EU AI Act Article 50 | AI-generated or manipulated content, or a deepfake, shown to EU users | Machine-readable mark; deepfake disclosure | Up to €15 million or 3% of global turnover |
The strictest applicable rule, not the platform minimum, should set the disclosure standard for any ad running across channels or jurisdictions. A campaign with an AI-generated presenter that runs on TikTok and reaches New York needs the TikTok AI-generated label, the ad-level AI Disclosure tag, and New York's conspicuous in-ad disclosure: three obligations, three mechanisms.
How do you build an AI disclosure compliance program?
A compliance program sorts content by risk, assigns a disclosure protocol per platform, writes AI duties into creator contracts, and keeps records. A brand that runs one creative on Meta, TikTok, YouTube, and X needs four disclosure executions, not one.
| Step | Action | Output |
|---|
| Categorize content | High risk: photorealistic people, AI endorsements, synthetic voices. Lower risk: AI-assisted writing and graphic elements. | A written risk framework |
| Set platform protocols | Write the disclosure steps for each platform | One checklist per platform |
| Cover creators | State who is responsible when a creator uses AI on brand content | Contract language and brief |
| Document decisions | Record AI production and disclosure choices | An audit trail |
| Review high-risk content | Send high-risk categories to counsel; add an election-cycle review | Sign-off record |
Watch these categories: AI-generated executive voices or images, AI-altered customer testimonials, AI product imagery presented as real photography, AI-generated reviews, AI-altered public figure content, and AI political content.
What is the decision for brands?
Brands using AI content at scale should audit current content against each platform's rules, New York and EU law, and the FTC's guidance, then build a structured program for the categories that produce volume. The exposure is real and the rules move, so a documented process protects a brand better than a post-by-post decision.
AI content disclosure FAQ
Do I need to disclose ChatGPT-generated content?
It depends on the use. Drafted blog or social text generally does not trigger platform disclosure. ChatGPT content used in endorsements, reviews, or testimonials that imply a human author or experience can trigger FTC disclosure duties. The commercial context decides, not the tool.
Does AI editing require disclosure?
Light editing such as color correction, audio cleanup, background removal, or grammar fixes generally does not. Heavier alteration, such as synthesized voices, AI faces, or fabricated scenes, typically does, and platforms ask whether the change would mislead a reasonable viewer.
Do AI-generated influencers have to disclose sponsorships?
Yes. Virtual influencers carry the same material-connection disclosure duties as human creators, and the AI nature of the influencer may need disclosure as well. That means two disclosure layers.
Do ads need a different AI disclosure than organic posts?
Often yes. On TikTok, ads carry a separate ad-level AI Disclosure tag on top of the AI-generated label, Meta applies automated labels with extra detection for political ads, and Google handles ad disclosure through My Ad Center and an asset-level control.
Can the FTC fine companies for undisclosed AI content?
The FTC enforces under existing authority, including Section 5 of the FTC Act and the Endorsement Guides, and outcomes have included settlements, injunctions, and consent orders. It has no dedicated AI disclosure rule.
Who is liable if a creator uses AI without disclosure on brand content?
Both brand and creator can face exposure, because the 2023 Endorsement Guides place a monitoring duty on brands. Put the disclosure requirement in the contract and the brief.
Does the EU AI Act apply to U.S. brands?
It applies when content reaches EU audiences, so a U.S. brand running paid or organic content in Europe needs to plan for it.
Which platform has the strictest AI disclosure rules?
Meta and YouTube enforce the most comprehensive requirements, with platform-level detection and labels, and TikTok adds a separate ad-level tag. LinkedIn is the most permissive of the major platforms.
