Do AI-generated ads need a disclosure?

Sometimes on the platforms, and always under deception law. Meta and Google each have an explicit advertiser disclosure rule, and both apply to political, election or social issue ads that use realistic synthetic people, events or audio. Outside those categories, the platforms we reviewed do not ask the advertiser to tick a box, but Meta may label the ad itself, and the FTC’s rules on endorsements and reviews apply to every ad that runs in the US.

This is a practitioner’s summary, not legal advice. For anything high-stakes (health claims, financial products, political work), have counsel review the ad.

What does Meta require?

Meta has two separate mechanisms: a disclosure rule for one category of ads, and labels that can apply to any ad.

The disclosure rule. Meta’s policy on ads about social issues, elections or politics requires advertisers to disclose when such an ad contains a photorealistic image or video, or realistic-sounding audio, that was digitally created or altered to:

  • show a real person saying or doing something they did not say or do,
  • show a realistic-looking person who doesn’t exist, or a realistic-looking event that didn’t happen, or alter footage of a real event, or
  • depict a realistic event that allegedly occurred but is not a true recording of it.

If an advertiser fails to disclose, Meta rejects the ad. Repeated failures may bring penalties against the advertiser. As of October 2026, the policy also says that beginning June 1, 2026, Meta uses automated detection to identify ad media created or edited with third-party generative AI, with no advertiser action required, and shows an “AI Info” label under About this ad.

Labels on other ads. Meta’s help page on AI-generated images in ads says ad images created or significantly edited with generative AI may be labeled, whether made with Meta’s own tools or third-party tools. Examples of significant edits include background generation and image generation. Minor changes such as resizing or color correction generally are not labeled. The label appears on the About this ad screen in the three-dot menu, and sometimes next to the Sponsored label. Meta says the labeling is rolling out gradually.

This matters for anyone using Meta’s own features. Several Advantage+ creative enhancements are generative AI (image generation, background generation, text generation, video expansion, AI music), and some enhancements may be on by default. Meta’s page also notes that certain regulated verticals, such as financial services, health, housing, employment and credit, and political ads, may not have immediate access to all generative AI features.

What does Google require?

Google’s explicit rule is for election ads. Its political content policy requires advertisers to disclose election ads containing synthetic or digitally altered content that inauthentically depicts real or realistic-looking people or events. The steps:

  1. In campaign settings, select the checkbox in the Altered or synthetic content section.
  2. For Feeds and Shorts on mobile phones, and in-stream placements on computers, mobile, mobile web and TV, Google generates the in-ad disclosure from that checkbox.
  3. For all other formats, add your own prominent, clear and conspicuous disclosure, such as “This image does not depict real events.”

Edits that are inconsequential to the ad’s claims are out of scope: resizing, cropping, color or brightness correction, red-eye removal, and background edits that don’t create realistic depictions of actual events. Google says a warning comes at least seven days before any account suspension under this requirement.

For all other ads, the relevant rule is Misrepresentation. It prohibits “manipulating media to deceive, defraud, or mislead others.” It doesn’t name AI, and it doesn’t need to: a synthetic before-and-after or a cloned voice implying an endorsement is manipulated media. Be careful with adjacent sections. Google treats Unacceptable business practices (which includes impersonating other brands or businesses) and Coordinated deceptive practices as egregious, with suspension on detection and without prior warning.

What does the FTC require?

The FTC doesn’t regulate “AI” as a category here. It regulates deception, and AI makes some deceptions cheap. Two rules matter most for performance advertisers.

The Endorsement Guides (16 CFR Part 255). An endorser “could be or appear to be an individual, group, or institution”, which reads naturally onto an AI-generated spokesperson. Three provisions do the work:

  • Actual consumers. If an ad presents endorsers as actual consumers, it should use actual consumers, or clearly and conspicuously disclose that they are not. This applies to hired actors and, by the same logic, to synthetic people.
  • Typical results. If the endorser’s results aren’t what people generally achieve, the ad should clearly and conspicuously disclose the generally expected results, and you need substantiation for that.
  • Substantiation. Claims made through an endorsement need the same support as if you made them directly. Consumer endorsements are not themselves competent and reliable scientific evidence.

In social media and other interactive media, the Guides say a disclosure should be unavoidable. The FTC’s Endorsement Guides Q&A adds that advertisers are responsible for what others do on their behalf, and that handing part of a promotional program to an outside company does not change that.

The Consumer Reviews and Testimonials Rule (16 CFR Part 465). In effect since October 21, 2024, it makes it a violation to write, create or sell a review or testimonial that misrepresents that the reviewer exists, that they used the product, or what their experience was. It also covers buying such reviews and disseminating testimonials the business knew or should have known were false. The FTC’s announcement names AI-generated fake reviews as an example and notes the rule allows civil penalties against knowing violators.

The FTC’s staff Q&A on the rule is specific about avatars: there is no blanket ban on AI avatars, a stock avatar’s message is prohibited under the rule only if the underlying testimonial is fake or false, using an avatar can still be deceptive under the FTC Act, and using a celebrity’s likeness without permission violates the rule if people would think the celebrity gave the testimonial.

Which rule applies to your ad?

Creative Meta Google FTC
AI product background or resized image, ordinary product ad May carry an AI info label for significant edits; no advertiser disclosure rule Misrepresentation if it misstates the product Product must look as it really does
AI avatar reading a real customer’s verified review AI info label possible Misrepresentation if misleading Disclose it is not the actual customer if presented as one
AI “customer” with an invented testimonial Label possible Manipulated media Prohibited fake testimonial under 16 CFR 465.2
Realistic synthetic person or event in an election or social issue ad Disclosure required; rejected if missing Altered or synthetic content checkbox plus disclosure Deception rules still apply
AI likeness or voice of a real celebrity without consent Label possible Misrepresentation Violates the reviews rule if presented as their testimonial

What should you check before launch?

  1. Is any person, voice or event in the ad synthetic? List each one.
  2. Is the ad political, electoral or about a social issue? If yes, disclose on Meta and check the Google box, and confirm the advertiser is authorized for that category.
  3. Does any synthetic person present as a customer? If yes, the words must come from a real, documented customer, and the ad should say the speaker is not that customer.
  4. Do the results shown reflect what people generally get? If not, disclose the generally expected results.
  5. Review every Advantage+ creative preview. Turn off enhancements that change claims you can’t support.
  6. Keep records: the source review, the consent, the prompt and tool used.

If you test AI variations against human-made ones, run the comparison cleanly; see how to test Meta ad creative. Landing pages need the same honesty as the ad; see landing pages for paid traffic.

How MSG approaches this

Clear, compliant advertising is one of our principles: no cloaking and no review evasion. When we produce performance creative, mostly short-form video, every synthetic element is listed before launch, testimonials trace back to real customers, and platform disclosures are applied where they are required. Creative, media and landing pages are reviewed together so the ad and the page make the same claims. For partners who pay on performance, the same standards apply to every traffic source we buy. See performance creative.

Sources

  1. Meta Transparency Center: Ads about social issues, elections or politics
  2. Meta Help Center: How AI-generated images in ads are identified and labeled
  3. Meta Business Help Center: About Advantage+ creative
  4. Google Ads policy: Political content
  5. Google Ads policy: Misrepresentation
  6. eCFR: 16 CFR Part 255, Guides Concerning the Use of Endorsements and Testimonials in Advertising
  7. eCFR: 16 CFR Part 465, Rule on the Use of Consumer Reviews and Testimonials
  8. FTC: Federal Trade Commission announces final rule banning fake reviews and testimonials
  9. FTC: The Consumer Reviews and Testimonials Rule, questions and answers
  10. FTC: FTC's Endorsement Guides, what people are asking

Platform features and policies change. This guide reflects the sources above as of October 7, 2026.