Do you have to label AI pics and videos on TikTok, Instagram and YouTube?

The short rule: label what looks real

Every big platform is converging on the same idea. If a viewer could mistake your AI-made or AI-edited content for a real person, place or event, tell them. If it is obviously stylized, animated or fantasy, the rules are much looser. Labels exist so people aren't misled, not to punish creativity.

TikTok

TikTok requires creators to label AI-generated content that contains realistic images, audio or video, and encourages labeling content that has been completely generated or significantly edited by AI. You can switch on the AI-generated content label when you post, which shows as "creator labeled as AI-generated", or add clear context with text, a sticker or the caption.

• Content made only with TikTok's own AI effects is labeled automatically.
• Uploads that carry Content Credentials (C2PA metadata) may be labeled automatically too.
• Unlabeled realistic AI content that could mislead can be removed.

Instagram, Facebook and Threads

Meta shows an "AI info" label when it detects industry-standard signals that an image or video was made or edited with AI, or when you disclose it yourself while posting. Meta asks people to disclose when they share photorealistic video or realistic-sounding audio that was digitally created or altered, and may apply penalties if you don't. Ads have their own disclosure rules, especially for political and social issue ads.

YouTube and Shorts

YouTube requires creators to disclose realistic altered or synthetic content in the upload flow, meaning content a viewer could easily mistake for a real person, place, scene or event. Once you disclose, a label appears in the description, and for sensitive topics like news, health, elections or finance it can also show on the video itself. YouTube doesn't require disclosure for content that is clearly unrealistic, animated, uses special effects, or where AI only helped with production.

Does my Dromana clip need a label?

It depends on what the clip shows. A few practical examples:

• An obviously stylized Y2K or 80s restyle of your own selfie: usually no label is required, but labeling is always fine.
• A realistic animated photo of you blinking and smiling, like the ones made with make a photo move: label it on TikTok, and consider disclosing on YouTube and Instagram.
• An old family photo brought to life: label it; people might assume it is real footage.
• A realistic clip of a real place with events that didn't happen: label it everywhere.
• Any realistic content about news, politics, health or money: label it and be careful.

If you're unsure, label it. It costs nothing, protects your account, and most viewers respect creators who are upfront.

What never flies, label or not

• Using someone's face without their permission to embarrass or impersonate them.
• Fake footage of real public figures saying or doing things they didn't.
• Realistic AI content designed to mislead people about news or events.
• Sexualized or unsafe content involving minors, in any form.

These break platform rules and our own safety guidelines. A label doesn't make them okay. If you're worried about your own photos being misused, read whether AI photo apps are safe.

How to add labels without killing the vibe

• Use the built-in toggle when posting; it's quiet and doesn't cover your content.
• Add a short line in the caption like "animated with AI" or "AI restyle".
• Make it part of the hook: "what AI thinks I looked like in 2003".
• Put a small text tag in a corner of the first frame for realistic clips.

Plenty of AI trends do better when the AI is the point. "If I lived in the 80s" is a great post precisely because everyone knows it's a restyle. Try it with the 80s photo filter, or learn how to make a photo move.