You have just finished creating an explainer video. The narration uses AI voice, the b-roll is generated, and the text is created automatically. When uploading, YouTube asks about the use of AI. Answer "Yes" or "No"?
The answer depends on one thing: whether the viewers can be misled into thinking what they see is real. The three major platforms use similar criteria, with different details.
YouTube: it's about realism, not the tool
YouTube's help page (accessed November 2026) asks creators to label content as realistic if it:
- makes real people appear to say or do something they never did,
- alters footage of real events or places,
- generates realistic scenes that never happened,
- uses AI-generated music as the main focus of the video.
YouTube also mentions examples that do not need to be labeled. Content that is clearly fantasy, beauty filters, color and light corrections, background blur effects, or vintage effects do not require a label. Likewise, AI used for drafting outlines, scripts, titles, thumbnails, infographics, and captions.
The process: when uploading in YouTube Studio, go to the attributes section, then select "Yes" for the AI usage question. The label may appear in the player or in the description. YouTube can also apply automatic labels for content created by its own AI tools and content with C2PA metadata. This type of automatic label cannot be removed.
Creators who repeatedly fail to label may receive a manual label from YouTube. Other sanctions include video removal or suspension from the YouTube Partner Program.
TikTok: labels for realistic AI
According to TikTok Newsroom, creators must label AI content that contains realistic images, audio, or video. AI-generated content is defined as content that is fully created by AI or significantly altered by AI. You can use the label button when posting, or add stickers and captions.
TikTok applies automatic labels for content created by TikTok's AI effects. Since May 2024, TikTok also reads Content Credentials (C2PA) from content generated by some other platforms. Therefore, videos from generators that embed metadata can be labeled directly. One strict limit: AI content that deceives and harms remains prohibited, labeled or not.
Instagram and Facebook: "AI info" label
Meta requires users to use the disclosure feature when posting photorealistic videos or audio that sounds real, created or altered digitally. Meta states that sanctions may be imposed if this obligation is ignored.
The label is called "AI info," a new name since July 2024 after the old label was applied to minor edits. Since September 2024, content that is only edited with AI tools will display that label in the posting menu. Content detected as AI-generated will still show an immediate label. For high-risk content that may mislead the public, Meta can apply a more prominent label.
Quick map for explainer videos and motion graphics
This table summarizes the rules above. The rows marked "explicitly mentioned" are quoted from the help pages; the rest are the result of applying the realism criteria of the three platforms.
| Use of AI | Need a label? |
|---|---|
| Script or outline from AI | No (explicitly mentioned by YouTube) |
| Automatic captions | No (explicitly mentioned by YouTube) |
| Color correction, face filters, background blur | No (explicitly mentioned by YouTube) |
| Removing background behind presenter | Generally no, unless the new background makes it seem like you are in a different real place |
| Motion graphics, illustrations, cartoon animations | No, as long as it is clear they are not real footage |
| Generic AI narrator voice | Depends on the platform; in Meta, audio that sounds real is also required to be disclosed |
| AI voice mimicking real people | Yes |
| Realistic human-faced AI avatar | Yes, because the realistic scene never happened |
| Photorealistic AI b-roll (city, office, crowd) | Yes |
| AI music as the focus of the video | Yes (YouTube) |
If in doubt, choose to label. The label does not change the content of your video, and viewers do not feel deceived when they find out later. The policies of the three platforms are also continually shifting, so check their help pages again before starting a new video series.
Sources
- YouTube Help: Disclosing use of altered or synthetic content
- TikTok Newsroom: New labels for disclosing AI-generated content
- TikTok Newsroom: Partnering with our industry to advance AI transparency and literacy
- TikTok Newsroom: More ways to spot, shape and understand AI-generated content
- Meta: Labeling AI-Generated Images on Facebook, Instagram and Threads
- Meta: Our Approach to Labeling AI-Generated Content and Manipulated Media