Calabi Labs · Guide · 2026-06-05
Ai disclosure rules by platform youtube instagramfacebook and tiktok l
AI Disclosure Rules by Platform: YouTube, Instagram, Facebook, and TikTok
If you're creating content on social media, here's what you need to know about each platform's AI disclosure requirements in 2024 and 2025.
YouTube
You must disclose AI-generated or altered content that looks realistic.
YouTube rolled out mandatory AI disclosure in March 2024. Creators are required to use the "AI-generated content" disclosure in YouTube Studio for any video containing realistic-looking content made or significantly altered by AI.
When disclosure is required:
- AI voice clones of a real person
- Digitally altered footage of real events or people
- Realistic fictional scenarios presented as real
- Photorealistic people or places that don't exist
When disclosure is NOT required:
- Clearly unrealistic or animated content
- Content using special effects (filters, color grading)
- AI used for background editing assistance without realistic simulation
- Scriptwriting, brainstorming, or metadata generation
YouTube uses automated detection and C2PA (Coalition for Content Provenance and Authenticity) standards. Content may be labeled "Altered" or "Synthetic" for viewers. Non-compliance can result in content removal or channel penalties.
Meta (Facebook and Instagram)
Meta requires AI content labels on both platforms, rolling out since May 2024.
Meta moved from a removal-based approach to a labeling model. Instead of deleting manipulated media, Meta now attaches "AI info" labels to organic AI-generated content across Facebook and Instagram.
What's labeled:
- AI-generated or AI-altered media that could mislead viewers
- Content produced or significantly modified by AI tools
Special rules for political ads:
- Any AI-generated content in political advertisements must be clearly disclosed
- This applies regardless of whether the AI content is the main subject or a supporting element
Meta uses C2PA and IPTC metadata standards to detect and label AI content automatically where possible. Creators should still manually label content when automated detection doesn't catch it.
TikTok
TikTok requires creators to label all realistic AI-generated content.
TikTok's AI labeling policy asks creators to use an in-app toggle to mark content that was wholly generated or significantly edited using AI. In May 2024, TikTok announced more proactive labeling, expanding beyond content made with TikTok's own AI effects.
When disclosure is required:
- Realistic AI-generated images or video
- AI-altered audio that sounds real
- Synthetic footage depicting real people, places, or events
Key distinction: Even properly labeled AI content can still be removed if it violates Community Guidelines. TikTok doesn't allow fake content that misleads viewers, even with a label.
TikTok also began automatically labeling AI-generated content from external AI tools in May 2024, using C2PA metadata when available.
Quick Comparison Table
| Platform | Disclosure Required | Automatic Labeling | Key Trigger |
|---|---|---|---|
| YouTube | Yes (mandatory) | Yes (where detectable) | Realistic-looking AI content |
| Yes (required) | Yes (gradual rollout) | AI content that could mislead | |
| Yes (required) | Yes (gradual rollout) | AI content that could mislead | |
| TikTok | Yes (required) | Yes (expanding) | Realistic AI images, audio, video |
What "Realistic" Means Across Platforms
All four platforms focus on content that could deceive viewers into believing something is real when it isn't. The threshold is content that looks or sounds genuine rather than obviously artificial or stylized. Animated content, clear visual effects, and purely creative AI assistance typically fall below the disclosure threshold.
The Technical Standard: C2PA
YouTube, Meta, and TikTok all increasingly rely on C2PA (Coalition for Content Provenance and Authenticity) metadata to detect AI-generated content. This technical standard embeds cryptographic provenance information into media files, allowing platforms to verify whether content was created or modified by AI tools. Content signed with C2PA credentials is easier for platforms to automatically label.
Bottom Line for Creators
- Always assume you need to disclose if your content shows realistic people, places, events, or audio that could be mistaken for real footage.
- Use platform tools — YouTube Studio, Meta's labels, TikTok's toggle — to mark AI content proactively.
- Don't rely on automatic detection — platforms may catch some content but not all; manual disclosure keeps you compliant.
- AI labels don't grant immunity — labeled content can still be removed for violating community standards.
Try Calabi free at calabilabs.com — 10 cleans, no card.