Do You Have to Label AI Ads? Rules for Meta, TikTok and the EU AI Act

What the EU AI Act requires

The EU's AI regulation (the AI Act) sets out transparency obligations in Article 50. For advertisers, one rule matters most: anyone who uses AI to generate or alter image, audio or video content that closely resembles real people, places or events must disclose that the content was artificially generated. These obligations have applied since August 2, 2026.

For ads, that means a UGC video with an AI person who looks like a real woman in her real car very likely falls under the rule. A clearly drawn illustration or an abstract background probably doesn't. The exact boundary hasn't been fully settled yet, so the safe practice is to always label realistic AI content.

If you only advertise in the US or the UK, the AI Act doesn't apply to you directly. Platform rules and consumer protection law still do, which is why labeling realistic AI content is the safe default in every market.

What consumer protection law has always required

Long before the AI Act, consumer protection law in the EU and the UK, and the FTC's rules in the US, prohibited misleading advertising. Two clear rules follow from that for AI ads:

• No fake testimonials: An AI person who says “I've been taking this for four months” must not be presented as a real, independent customer experience. Fake reviews and testimonials are explicitly prohibited.
• No invented experts: An AI doctor in a white coat recommending your product is doubly problematic: it's misleading, and for health products it also runs into the stricter rules for health advertising.

Meta's rules

Meta labels content with “AI info” when it detects it as AI-generated, for example through technical signals in the file, or when advertisers disclose it themselves. For ads about social issues, elections or politics, Meta explicitly requires disclosure when realistic content has been digitally created or altered. For regular product ads, the rules are less strict, but the direction is clear: transparency is expected, and misleading content violates Meta's advertising standards.

TikTok's rules

TikTok requires realistic-looking AI content to be labeled and offers its own label for it. Content that deceptively imitates real people or misleads viewers about its authenticity isn't allowed. When you upload ads, check whether the option to label AI content is offered, and use it.

What clean labeling looks like

1. In the image or video: a small, clearly legible note such as “Made with AI,” for example in the bottom right corner. At AVVEN, it's already built into every ad that features AI people.
2. In the caption: an additional short sentence or a note like “AI-generated.”
3. On the platform: use the platform's own AI label or toggle if there is one.
4. In the script: no phrasing that claims a real experience, such as “as a customer for three years,” when the person is artificial.

What you can safely do with AI

• Product scenes, backgrounds and lighting moods
• Illustrations, icons and graphic elements
• Variations of copy, hooks and layouts
• AI people in a clearly staged ad scene, with a label

Labeling and performance

Many brands worry that a “Made with AI” note will hurt results. In practice, the hook, the problem and the offer decide whether an ad succeeds, not a small note in the corner. It's far more expensive when ads get rejected, an ad account gets restricted or customers feel deceived.

This article is not legal advice and reflects the situation as of October 2026. Platform rules change regularly. For specific questions, talk to a qualified attorney who specializes in technology or advertising law.

Frequently asked questions