Do you have to label AI ads?
Meta, TikTok, Google and YouTube rules

Do you have to label AI-generated ads?

It depends on the platform, and the rules changed this year:

TikTok: label it or the ad is rejected or restricted

TikTok allows content “changed by AI beyond minor tweaks or enhancements” in ads on one condition: “Apply the AIGC label, or by adding a clear disclaimer, caption, watermark, or sticker of your own.” Without it, the rule is blunt: “If we identify AI-generated content that has not been disclosed, your ad will be rejected or restricted.”

Its examples: “Content that contains images, video, or audio that are completely AI-generated”, “Showing the primary subject doing something they didn’t actually do”, or “Making the primary subject say something they didn’t actually say, using AI voice-cloning.” Minor fixes such as lighting, colour or “Removing or modifying backgrounds” need no label. In doubt, TikTok says, label it. Ads Manager lists “AI-generated, synthetic, or manipulated media” under mandatory disclaimers.

Meta: Facebook and Instagram label it for you

Meta says it “will automatically label ad content that has been created or edited using some of Meta’s generative AI tools or with third-party generative AI tools”. It finds outside tools through “industry-standard detection methods, such as C2PA”, the content provenance standard. Not every ad gets labelled yet: Meta says that while the tools roll out, “not all ads using Meta’s generative AI creative features will have AI info”.

The label sits under “About this ad” in the three-dot menu, and sometimes at the top of the ad. Minor changes such as resizing or colour correction are not labelled, but using Meta’s Background Generation, Image Generation or Add Animation features is. Meta updated these labels in July 2026, Social Media Today reported.

You must act yourself on ads about social issues, elections or politics, which must disclose “a photorealistic image, video or realistic-sounding audio that was created or edited using AI tools.” Where local law asks for AI transparency, Meta also lets advertisers self-disclose; it names “the European Region, California, New York, India and Taiwan.”

Google and YouTube ads: the new AI label setting

Since July 2026, advertisers can mark assets as made or edited with AI, shown in a new “How this ad was made” panel on Search, YouTube and Discover ads, Google announced. Its policy update gives the reason: “AI regulations in the European Union, India, and New York require that ads with certain AI-generated or edited assets include disclosures and/or labels”.

For campaigns that target those three places, ads marked as AI-made “will also include visible overlays on the ads”. Your own label inside the creative is now allowed too: it “won’t be in violation of Google policies prohibiting text overlays and watermarks.”

Election ads that contain synthetic or digitally altered content must still be disclosed with the “Altered or synthetic content” checkbox, and Google warns that its setting “doesn’t guarantee your compliance with specific regulations.”

YouTube uploads: disclose what looks real

For videos you upload, YouTube’s rule is: “We require creators to disclose AI-generated or altered content that seems realistic.” One of its examples reads like an ad brief: “AI generated extra footage of a real place, like a video of a surfer in Maui for a promotional travel video.” Clearly unrealistic content and minor edits need no label, and neither does “Cloning one’s own voice to create voice overs or dubs”.

YouTube may add a label itself to content with C2PA metadata, and says: “Disclosing AI content alone won’t limit a video’s audience or impact its eligibility to earn money.”

Does an AI label hurt the ad?

None of the ad policy pages we read says a label lowers delivery. Ipsos reports that 79% say companies using AI should have to disclose it, and that brands caught using AI without disclosure “have faced swift vocal backlash”.

A checklist before you launch

• Plan for the label. For realistic AI people, places or voices, TikTok and YouTube uploads require one; Meta and Google add one in some cases.
• Expect detection. Meta and YouTube read C2PA metadata, so a label can appear whatever you tick.
• Never borrow a famous face or voice. TikTok does not allow misuse of a public figure’s likeness “without their permission”.
• Check the country. In the EU, India and New York, Google says the rules ask for disclosures or labels, not only the platform.

We follow the rules of the platform each ad runs on and tell you when a label is needed, in every AI video ad we make.

This is general information, not legal advice.