How to Label an AI-Generated Video Ad

There's no single “AI-generated” sticker that satisfies every platform, because each one has a different rule. TikTok requires its AIGC label (or your own clear disclaimer, caption, or watermark) on realistic AI-made ad content, or the ad gets rejected. Meta shows an “AI info” label that applies automatically when you use its own generative tools, or that Meta now detects on third-party AI content and applies itself. The FTC doesn't require any specific label at all — it requires that whatever the ad shows, and whoever appears to say it, isn't a lie. Flovaly doesn't burn in a disclosure for you, so treating it as a deliberate step in your workflow, not an afterthought, is what keeps an ad compliant everywhere it runs.

Why “just add an AI label” isn't one task

It's tempting to treat AI disclosure as a single checkbox, but the three bodies that actually govern it — TikTok, Meta, and the FTC — aren't coordinated and aren't asking for the same thing. TikTok and Meta are platform policies: break them and your ad gets rejected or restricted, nothing more. The FTC is a federal regulator applying general truth-in-advertising law: break that and you're looking at deception enforcement, independent of whatever label sat on the post. Clearing one doesn't clear the others, so it's worth knowing what each actually requires.

TikTok: the AIGC label, or your own clear disclaimer

TikTok's advertising policy is explicit: ads containing content that's completely AI-generated, or “significantly modified” by AI — its own examples include making a person appear to do something they didn't do, or say something they didn't say via AI voice-cloning — need to be disclosed by applying “the AIGC label, or by adding a clear disclaimer, caption, watermark, or sticker of your own”. TikTok says plainly that if it identifies undisclosed AI-generated content, “your ad will be rejected or restricted” — so this isn't a best-practice suggestion, it's a submission requirement enforced before the ad runs.

Meta: the “AI info” label, self-applied or detected

Meta's rule triggers when “an image or video is created or significantly edited with our generative AI creative features in our advertiser marketing tools,” or when the ad contains a photorealistic AI-made human — in Meta's own words. If you build the ad with Meta's in-house generative tools, the “AI info” label applies automatically. If you make the video elsewhere — on Flovaly, for instance — and upload it, Meta says it will “begin automatically detecting ads created or edited using third-party AI tools through industry-standard signals” and apply the label itself once detected. For a photorealistic human specifically, the label sits “next to the Sponsored label,” not tucked behind a menu — Meta treats that case as higher-stakes than a generic AI-touched product shot.

The FTC: no label requirement, just no lying

Search results are full of claims about a “2026 FTC AI disclosure law.” There isn't one. What actually exists is the FTC's longstanding Section 5 authority over deceptive advertising, applied to AI content the same way it's applied to anything else. Its Endorsement Guides require that “an endorsement must reflect the honest opinion of the endorser” — a rule that predates AI video entirely and doesn't care whether the endorser was filmed or generated, only whether the opinion is real. Separately, the FTC's August 2024 final rule banning fake reviews and testimonials, effective since October 2024, bans manufacturing reviews or testimonials that don't reflect a real customer's experience, with civil penalties available against knowing violators. Neither rule asks for a label on the video. Both ask that an AI presenter not be used to fabricate an endorsement, result, or customer that doesn't exist.

How to actually disclose it on a Flovaly-made ad

Flovaly doesn't add an “AI-generated” watermark or label to Talking Actors, UGC, or Re-voice clips — nothing in the generation pipeline stamps one on, so the disclosure step is on you, done where each platform actually looks for it:

• Uploading to TikTok Ads Manager— apply the native AIGC label at the ad level. If you'd rather not rely on the platform toggle alone, write the disclosure straight into your script or on-screen text, since Talking Actors and UGC clips already burn in captions by default — a line like “AI-generated video” sitting in that same caption layer counts as TikTok's “clear disclaimer, caption, watermark, or sticker” alternative.
• Uploading to Meta Ads Manager— if the ad features a photorealistic AI actor, don't assume automatic detection catches it before submission; check for a self-disclosure option in Ads Manager and use it. Meta's own language treats an undetected photorealistic AI human as the highest-scrutiny case, not the safest one to leave unlabelled.
• Anywhere the ad runs as a customer testimonial— add a plain on-screen line such as “real customer review, read by an AI presenter”, the same disclosure pattern covered in making an honest AI testimonial video. That satisfies the FTC's honesty requirement regardless of which platform the clip ends up on.

When it's genuinely lower-stakes

Not every AI-touched ad needs the same scrutiny. Meta's own wording carves out tools that “do not result in significant edits to the image or video and do not include a photorealistic human” — a product shot with a light AI colour-correction pass isn't what either platform's policy is aimed at. Where an ad is clearly stylised or obviously synthetic rather than presented as unedited footage of a real event, the deception risk the FTC cares about is lower too. The clearest trigger across all three — TikTok, Meta, and the FTC alike — is a realistic AI human appearing to say or do something a real person didn't. That's the case to always disclose, however light the edit otherwise looks.

None of these rules change what Flovaly can generate — they change what you do with the finished clip before it goes live. Build the ad, then spend the extra minute labelling it correctly for wherever it's running. Start with the $1 trial, or read how to weigh an AI actor against hiring a real UGC creator if disclosure risk is part of that decision for you.

FAQ

Do I have to label an AI-generated video ad?

It depends where it runs. TikTok and Meta both require disclosure on realistic AI-generated ad content as a platform policy — break it and the ad gets rejected or restricted. The FTC doesn't require a label at all; it requires that the ad not misrepresent a real person's endorsement or experience, label or no label.

Does TikTok require an AI label on ads?

Yes, for ads with completely AI-generated or significantly AI-modified content. TikTok's advertising policy requires applying its AIGC label, or your own clear disclaimer, caption, watermark, or sticker — undisclosed AI content that TikTok identifies gets the ad rejected or restricted.

Does Meta require an AI label on ads?

Yes, when an ad is created or significantly edited with generative AI, especially if it includes a photorealistic human. Ads built with Meta's own AI tools get the “AI info” label automatically; ads made elsewhere and uploaded may get it applied once Meta's detection catches the AI signals, but self-disclosing rather than relying on detection is the safer default for a photorealistic AI actor.

Does the FTC require an “AI-generated” label?

No — there's no FTC rule requiring a specific AI label. Its Endorsement Guides and its 2024 rule banning fake reviews and testimonials both apply regardless of a label: an AI-voiced endorsement still has to reflect a real person's actual opinion, and a fabricated customer or result is prohibited whether or not the video says “AI-generated” on it.

Does Flovaly add an AI-generated label to my videos automatically?

No. Flovaly doesn't burn in a watermark or disclosure label on Talking Actors, UGC, or Re-voice clips, so applying TikTok's or Meta's native ad-level disclosure, or writing your own on-screen line, is a step you take after the clip is generated — not something the Studio handles for you.

FAQ