Is AI Content Labeled on Meta? What Advertisers Should Know

Yes, Meta requires advertisers to disclose when ads use AI to create or alter content in ways that could mislead viewers, particularly around realistic depictions of people, voices, or events, and it has built-in disclosure tools for this in Ads Manager. If your ad uses an AI-generated actor delivering a script or a synthetic voiceover that could be mistaken for a real unaltered recording, plan to disclose it. Policy specifics and enforcement mechanics change over time, so check Meta's current advertising standards and AI disclosure requirements directly rather than relying on a summary.

Why should advertisers take this seriously?

Undisclosed content that trips Meta's policy can get an ad rejected, delayed in review, or in repeated cases put the ad account's standing at risk - all more costly than the small step of applying a disclosure label. There's also a trust dimension beyond compliance: viewers are getting better at spotting AI content, and disclosing it tends to preserve credibility better than being caught concealing it.

What triggers a disclosure requirement?

Broadly, realistic AI-generated or substantially AI-altered content depicting real or real-seeming people, places, or events - this is the category Meta focuses on, as opposed to obviously stylized or clearly synthetic-looking content. If a viewer could reasonably mistake the ad for unedited real footage, that's the signal to disclose.

Does disclosure change the creative itself?

Not fundamentally. Whether or not the ad carries a disclosure label, the same UGC ad fundamentals still apply: vertical framing for feed and Reels placements, a hook that lands in the first couple seconds, and key visual elements kept clear of platform UI overlays. Disclosure is a layer on top of the creative, not a redesign of it.

How should this factor into an AI ad testing workflow?

Build the disclosure check into your pre-launch checklist the same way you'd check aspect ratio or caption placement, rather than treating it as an afterthought. Because the production side of AI UGC ads is now fast - describe the ad in Claude with Polaris connected, attach a product photo or script, and a live panel renders the finished video in the same conversation - the disclosure step is a small addition, not a workflow overhaul. From there, measure disclosed AI creative the same way you'd measure any variant: watch hook rate and downstream ROAS using the hook generator and ROAS calculator. More on the production workflow is in generating ads inside Claude via MCP and the AI UGC ads overview.