AI Ads On Meta & TikTok: Rules, Labels And What Gets Rejected

AI makes ad production faster than ever — but Meta (Facebook and Instagram) and TikTok are watching AI-generated content closely. A rejected ad costs you time; a restricted ad account costs you a lot more. This guide covers what the platforms expect from AI ads and how to stay on the safe side.

Note: platform rules change regularly. Treat this as a practical overview and always check the current Meta Advertising Standards and TikTok Advertising Policies before launching.

The principle behind every rule: don’t mislead

Both platforms allow AI-generated content in ads. What they don’t allow is content that deceives people. If an AI element makes viewers believe something false — about who is speaking, what the product does or what results to expect — you’re in risky territory.

Meta (Facebook & Instagram)

• Disclosure for sensitive topics. Ads about social issues, elections or politics that contain realistic, digitally created or altered images, video or audio must be disclosed when you set them up. Most e-commerce ads don’t fall into this category — but know that it exists.
• “AI info” labels. Meta labels content it detects as AI-generated or edited. It doesn’t block the ad, but viewers can see it — another reason not to pass AI off as real.
• No fake endorsements. Using a public figure’s face or voice — real or AI-generated — without permission is a fast way to get rejected and restricted.
• Health and body claims. Ads for weight loss, skin and health products face strict rules on before-and-after images and implied results. An AI-generated “after” photo is both against the rules and simply misleading.

TikTok

• Label realistic AI content. TikTok expects realistic AI-generated content to be clearly disclosed. For ads, Ads Manager provides a way to declare that an ad contains AI-generated content — use it whenever it applies.
• No misleading likenesses. AI content that impersonates real people, especially public figures endorsing products, isn’t allowed.
• Authentic-looking claims still need to be true. A “creator” testimonial — real or AI — must not invent results or experiences the product can’t deliver.

What usually gets AI ads rejected

1. AI-generated before-and-after results, especially for health, weight or skin.
2. Avatars or voices that look like celebrities or influencers.
3. AI “customers” presented as real buyers with personal testimonials.
4. Product visuals that show features the real product doesn’t have.
5. Exaggerated claims in AI voiceovers (“cures”, “guaranteed”, “instant”).

Our checklist before an AI ad goes live

• Does every product shot match the real product — logo, colour, size, label?
• Are all results shown real and typical?
• Is it clear the ad isn’t impersonating anyone?
• Have we used the platform’s AI disclosure where it applies?
• Would we be comfortable if a customer asked, “Is this real?”

Use AI where it’s safe — and strong

The safest and most effective uses of AI are mood and product visuals, locations, ingredient animations, voiceovers and fast variations of a proven script. Keep proof and personal experiences real. Read How We Mix AI And Real Footage for how we combine the two, or see our AI ads. Not sure your concept is compliant? Ask us before you spend budget on it.