AI UGC ads on Meta in 2026: labels, policy, what gets approved.

Short answer: yes, you can run AI-generated UGC ads on Facebook and Instagram. Meta does not ban AI creative. It reviews an AI-made ad against the same Advertising Standards as a human-shot one, and it may attach an AI info label in the About this ad menu when the ad was created or significantly edited with AI. Self-disclosure is only mandatory for one category (ads about social issues, elections or politics). What gets AI UGC rejected is almost never the avatar. It is the claim the avatar makes.

Here are the rules as they read in August 2026, from Meta's own pages, with a short TikTok comparison. Policy changes; when in doubt, read Meta's current Advertising Standards before launch.

Start a brief→

The AI info label: what it is and when Meta applies it

Meta's help center says that if an ad was created or significantly edited using Meta's generative AI creative features, or created or edited using third-party AI tools, the AI info label appears on the About this ad screen, reached from the three-dot menu in the top right of the ad. In some cases it is placed beside the Sponsored label. Significant edits such as background generation and image generation trigger it; minor adjustments such as resizing and color correction do not. Meta also notes that as the tools roll out, not every ad made with its generative features will show the label yet (source).

The label itself came from Meta's 2024 organic-content work: "Made with AI" was renamed "AI info" in July 2024, and labeling is based on industry-standard signals embedded by AI tools plus self-disclosure (source). Practically: if your AI UGC ad carries standard provenance metadata, or you used a Meta generative feature, expect the label. It sits in a menu most people never open.

When disclosure is mandatory: one category

Self-disclosure is required for ads about social issues, elections or politics. If such an ad contains a photorealistic image or video, or realistic-sounding audio, that was digitally created or altered to depict a real person saying or doing something they did not, to depict a realistic-looking person who does not exist or event that did not happen, or to present a realistic event that allegedly occurred but is not a true recording, the advertiser must disclose it. Meta rejects the ad if they do not, and repeated failure can bring penalties against the advertiser (source).

This is category-specific. A DTC skincare or cookware brand is not in it. If your brand touches an advocacy, civic or policy topic, you are, and an AI UGC presenter is exactly the kind of realistic person who does not exist that the rule was written for.

Meta's own generative features

Advantage+ creative ships a set of generative options inside Ads Manager: background generation, image expansion, video expansion, image animation, text generation, image generation and image touch-up. Meta marks any feature that uses AI with an AI logo under the feature name (source). Using them is one of the named triggers for the AI info label, and they rewrite finished creative: expansion can stretch a product shot, text generation can reword a claim you already cleared. Turn off what you did not ask for and check the rendered placements, not just the upload.

What gets AI UGC rejected (it is the script, not the face)

Meta's Unacceptable Business Practices standard prohibits deceptive or exaggerated claims about the success of a product or its health-related benefits (source). AI UGC makes this easier to violate, because it is cheap to script a "customer" who says whatever converts. A synthetic person reciting results that did not happen is a deceptive claim with a face on it. We do not produce fake-consumer testimonials for that reason.

The Health and Wellness standard adds category rules. Ads for dietary, health, weight loss or weight gain products, and ads for cosmetic products, procedures or surgeries, must be targeted to people 18 or older. Ads cannot claim to cure, heal or eliminate a list of named conditions. They cannot use clickbait such as extreme claims or promised outcomes within a set timeframe without disclaimers, and they cannot contain statements of inferiority about someone's appearance. On before-and-after: Meta lists general cosmetic products, procedures and surgeries depicting a before and after transformation as permitted when targeted 18+; it does not extend that allowance to weight-loss transformations in the text we fetched (source). If your ad lives near those lines, check the page yourself before you commit a batch.

The general standards also require that every ad component be relevant to the product, and that the product on the landing page match the product in the ad (source). "Unrealistic results" in AI UGC usually shows up as a rendered scene doing something the product cannot, like a spray erasing a stain in one frame. Review the pixels for implied claims, not just the words.

TikTok, briefly

TikTok is more explicit. Ads Manager has an ad-level toggle that reads "This ad contains AI-generated content", meant for image, video or audio that is completely AI-generated or significantly modified by AI. Once submitted, you cannot change or unselect it unless you change the video or image, and any new campaign, including duplication, resets it (source). TikTok's ad policy also accepts a clear disclaimer, caption, watermark or sticker instead of the label, restricts before-and-after comparisons that could distort a product's outcome, and prohibits exaggerated effect claims (source). Build the toggle into your launch checklist; it is the step that gets lost when someone duplicates a winning campaign.