Do you have to label AI ads? Meta, TikTok and Google rules for Canadian businesses in 2026
Short answer: On TikTok, yes: ads with realistic AI generated people, voices or scenes need the AI disclosure switched on in Ads Manager, which puts a visible label on the ad. Meta can label AI content automatically and gives you a disclosure option in Ads Manager. Google Ads asks you to indicate generative AI use. In Canada, a misleading ad is illegal whether a person or an AI made it.
Most small business owners running their first AI video ad ask us the same nervous question: will Instagram slap a big "fake" sticker on it?
Short version, no. But the rules did tighten through 2026, and they are different on every platform. Here is what Meta, TikTok and Google actually ask for, what Canadian law adds on top, and the checklist we'd run before uploading anything.
This is a practical guide, not legal advice. Platform policies change often, so confirm against each platform's help centre before a big campaign.
Why AI ad labels became a thing in 2026
Two changes landed at roughly the same time. First, the platforms started reading content credentials (an industry standard called C2PA) that many AI tools now embed in their files. That lets Meta and TikTok spot AI generated media even when it was made somewhere else and uploaded later. Second, TikTok and Meta both updated their ad transparency rules during 2026, with Meta rolling out a single "About this ad" panel from June 1, 2026 that includes AI information.
So the label question is no longer "will anyone notice." Often the platform notices on its own. The question is whether you disclose properly and whether your ad is honest about what it shows.
What each platform requires
| Platform | What triggers a label | What you do | Where viewers see it |
|---|---|---|---|
| Meta (Facebook, Instagram, Threads) | Ads created or significantly edited with generative AI, including Meta's own creative tools and third party AI it detects | Use the AI disclosure option in Ads Manager; political and social issue ads must self disclose | An "AI info" note, mainly inside the "About this ad" panel |
| TikTok | Realistic AI generated or significantly edited people, voices, scenes or photorealistic product renders | Turn on the AI disclosure in TikTok Ads Manager before submitting | A visible label on the ad itself |
| Google Ads | Ads created or edited with generative AI, inside or outside Google's tools | Use the control for indicating generative AI use | "How this ad was made" information |
| YouTube | Realistic synthetic content that viewers could mistake for real | Creator side disclosure at upload | A label in the description or on the player |
TikTok is the strict one
If your ad shows a realistic AI person, a cloned or synthetic voice, an AI background that looks like a real place, or an AI rendered product that looks like a photo, TikTok wants the disclosure on. AI generated political ads are banned outright as paid ads. The upside: TikTok has said that switching on the AI content setting does not reduce distribution as long as the video follows its community guidelines.
Meta is more relaxed for ordinary commercial ads
For a bakery or a skincare brand, Meta mostly handles labelling through detection, and the note sits one tap away in the ad's info panel rather than across the video. That said, Meta has made clear the advertiser running the ad is responsible for AI variants in it, even ones generated by Meta's own tools. If you use the disclosure toggle yourself, you are covered either way.
What does not need a label
You do not need to disclose that AI helped you write the caption, the headline, the hashtags or the on screen text. Script help is also exempt on the major platforms. The rules are aimed at realistic visuals and audio, the things a viewer could mistake for real footage.
What Canadian law says about AI in ads
As of October 2026, we are not aware of any federal Canadian rule requiring a label on every AI generated ad. What does apply is the Competition Act, which makes false or misleading representations in advertising illegal, enforced by the Competition Bureau. AI doesn't change that test. It just makes it easier to fail.
Where AI ads get Canadian businesses into trouble:
- Fake testimonials. An AI person saying "I used this for three weeks and my skin cleared up" is presented as a real customer experience. It isn't one.
- Changing the product. If the AI makes your portion bigger, your fabric shinier or your results more dramatic than reality, that's a misleading representation.
- Invented claims. Anything the video implies (health effects, "made in Canada", a discount) still has to be true.
Selling into the US as well? The FTC expects two separate signals on paid content: that it is an ad, and that AI was used. An undisclosed synthetic endorsement is treated as a deceptive testimonial.
Does labelling an AI ad hurt performance?
From what the platforms say and what we've seen, a disclosure on its own isn't what sinks an ad. People scroll past an "AI info" note the same way they scroll past "Sponsored". What hurts is an ad that feels off: the morphing hands, the too smooth faces, the obviously stock AI presenter reading a script. Viewers react to that whether there's a label or not.
This is one reason we lean on lifestyle footage of the product in use rather than AI people pretending to be customers. It is easier to keep honest, and it doesn't ask anyone to believe a person exists who doesn't.
Checklist before you upload an AI video ad
- Does it show a realistic AI person, voice or place? If yes, switch on disclosure on TikTok and use Meta's AI option.
- Does any AI person appear to be a real customer giving a review? If yes, rewrite it.
- Does the product on screen match what someone actually receives? Size, colour, texture, portion.
- Is every claim in the video, captions and voiceover true and provable?
- Running political or social issue content? Different, stricter rules apply on every platform. Check before spending a dollar.
- Keep the original files and your brief. If an ad is ever questioned, you can show how it was made.
Frequently asked questions
Do I have to label AI generated ads on Instagram and Facebook?
Meta applies an AI info label automatically when it detects generative AI, and gives advertisers a disclosure option in Ads Manager. For ordinary commercial ads the note appears in the About this ad panel. Political and social issue ads must self disclose AI use.
Does TikTok require AI disclosure on ads?
Yes. TikTok requires advertisers to enable AI disclosure in Ads Manager when an ad contains realistic AI generated or significantly edited people, voices, scenes or photorealistic product images. This adds a visible label to the ad. Political AI ads are not allowed.
Is there a Canadian law that requires AI ads to be labelled?
As of October 2026 there is no federal Canadian rule requiring a label on every AI ad. The Competition Act still prohibits false or misleading advertising, so AI ads must accurately represent the product and must not present fake testimonials as real.
Do AI written captions need a disclosure?
No. On Meta and TikTok, AI help with captions, headlines, hashtags, text overlays and scripts does not require a label. The rules target realistic AI visuals and audio.
Will an AI label reduce my ad's reach?
TikTok has said enabling its AI content setting does not reduce distribution for content that follows its guidelines. On Meta the label sits in the ad info panel. Poor quality or misleading creative is a much bigger risk to performance than the label itself.
AI ads that stay honest
We start from your real product photo and build lifestyle scenes around it. Every ad is reviewed by a person before it reaches you. From $49 CAD.
See pricing