How Facebook shows the label
On Facebook the mark appears as “AI info” in the post menu or under the image, and it is shared across Meta’s ecosystem (Instagram, Threads). Meta applies the label using provenance standards such as C2PA, so embedded credentials weigh heavily in the automatic decision.
How to remove it, step by step
- Upload the image to the checker and review what it contains: C2PA, IPTC, EXIF or XMP.
- Clean the file or use the /c2pa mode to remove provenance credentials.
- Download the clean version and confirm again in the checker.
- Publish on Facebook without passing the image through another app that reinjects metadata.
Nuance and false positives
Real photos retouched with AI features, or exported from Photoshop, Lightroom or Canva with provenance on, can be marked even when they were not generated. Cleaning the metadata fixes that false positive. Even so, Facebook can detect by other means, and we do not act on invisible pixel-level watermarks.
Final tip
If you reuse an image you already published and edited, check it again: when you save it, many tools rewrite the metadata you had removed.