Why Meta Sometimes Puts an AI Label Right Next to 'Sponsored' – and Otherwise Hides It

Since June 1, 2026, Meta has expanded the transparency information on Facebook and Instagram ads, bundled inside "About this ad" in the three-dot menu: notices about AI-generated or AI-edited content, information on who pays for or benefits from an ad, and "Paid for by" labeling for political ads. Meta now also automatically detects AI content made with third-party tools, not just its own. Most of these notices stay tucked inside the menu, one click away. There's one exception: when an ad shows an AI-generated, photorealistic person, the label can appear right next to the Sponsored notice itself – prominent, visible without being asked for, not hidden behind a click.

What that one placement decision reveals

An interface decision like this is rarely an accident. That Meta specifically lowers the visibility threshold for synthetic people is the same kind of involuntary admission I already described regarding ChatGPT advertising: a company that labels one category of content more strictly than another reveals its own internal assessment of where the biggest risk of deception sits. An AI-optimized background image or an automatic color correction gets the quiet treatment, tucked in a menu. A face that looks real but doesn't exist gets the prominent one.

The difference from Article 50

It's worth not confusing these layers: Meta's placement decision is a voluntary product choice, not a legal requirement – unlike Article 50's narrow process boundary, which dictates when labeling is legally mandatory at all. In parallel, since August 18, 2026, the second version of the IAB Transparency Framework has existed as a global industry standard demanding disclosure specifically for photorealistic AI images – independent of EU law. Three different layers, then, that happen to land on the same point: law, industry standard, and a single platform voluntarily going further than either.

What this means for using synthetic models

For brands using synthetic models or digital twins in their advertising – as I already described using About You as an example – this placement decision isn't a footnote, it's a real creative constraint. An AI-generated model now visibly competes in the image with a notice sitting right next to the Sponsored tag, instead of quietly disappearing into a menu. That changes the math: anyone choosing synthetic faces because they're faster and cheaper has to factor in that this exact advantage gets made visible on the platform itself, before anyone has even consciously registered the image.

Conclusion

With a single placement decision, Meta has made explicit something that was already obvious but rarely stated outright: not every AI-generated element carries the same risk. A face that looks real but belongs to nobody gets the most prominent label of all – not because a law demands it, but because the platform itself considers it the most sensitive case.