What C2PA is, in one paragraph
C2PA (Coalition for Content Provenance and Authenticity) is an open standard for cryptographically signing metadata onto an image, video or audio file that describes how it was made and by what tools. When a video is generated or edited by a C2PA-aware tool, the tool embeds a signed manifest into the file that a downstream platform can inspect: which AI models touched the file, what edits were applied, whether the underlying assets were themselves AI-generated. Think of it as an audit trail that travels with the file, verifiable without trusting the sender.
Why platforms suddenly care
Two forces converged in 2024-2026. Regulators in the EU (AI Act, Article 50) and multiple US states now require AI-generated content to be disclosed when used in advertising or political communication. Simultaneously, ad platforms discovered that undetected synthetic content was eroding user trust in what they were seeing on the feed. C2PA gives platforms a technical mechanism to comply with regulation and rebuild trust at the same time — they can automatically label AI content instead of relying on self-disclosure that most advertisers were skipping.
Meta's requirements (2026)
Meta requires advertisers running political, election, social-issue or health ads to affirmatively disclose AI generation at upload — a mandatory checkbox in Ads Manager. For all other advertising, Meta's own C2PA detection pipeline scans uploaded videos and applies an "AI info" label automatically when synthetic content is detected. Ads that carry the label perform on par with non-labeled ads in Meta's own testing; ads that omit disclosure and are later flagged get throttled while the account is reviewed. Best practice: let the C2PA-embedded manifest do the disclosure work automatically at upload, rather than toggling the checkbox manually per ad.
TikTok's AI content policy
TikTok requires any AI-generated content that features realistic-looking people, places or events to be labeled — either creator-applied via the "AI-generated content" toggle in the upload flow, or auto-detected via provenance metadata (C2PA compatible). Ads that carry the label are treated as first-class inventory; unlabelled AI ads that TikTok's detection catches after the fact get removed and count as a policy violation against the ad account. The platform's Commercial Music Library and its Symphony creative tools are pre-integrated with C2PA, so first-party content flows through the pipeline correctly by default.
YouTube Shorts and the "altered content" label
YouTube requires disclosure of "meaningfully altered or synthetically created" content in a required upload field, and adds an "altered content" label to the video watch page automatically when disclosure is missed but detection catches it. For Shorts specifically, the label appears in the video description. YouTube's C2PA support arrived alongside Google's broader Content Credentials rollout in 2024-25; providing a signed manifest satisfies the disclosure requirement without needing per-video manual attestation.
How UGCABC handles this automatically
UGCABC exports do not yet carry a C2PA manifest; Content Credentials are on the roadmap, not in the file today. Until they are, use each platform's own AI-content toggle at upload — Meta, TikTok and YouTube all have one — and keep your own record of which ads are AI-made. The disclosure takes seconds; the throttling for a missed one takes weeks to recover from.
What still needs manual attention
C2PA covers the "is this AI" question; it doesn't cover the claim itself. If the ad makes a medical, financial or performance claim, you're still responsible for substantiation under the same rules that apply to human-produced ads. Regulated categories (crypto, health, political) still require the category-specific attestations each platform demands — C2PA doesn't exempt you from Meta's health-claim policies or FTC endorsement rules. Treat automated AI disclosure as a compliance floor, not a ceiling.