The short answer
Instagram shows "AI info" on a real photo when the file contains an AI marker, usually C2PA Content Credentials or an IPTC DigitalSourceType written by an app that used a generative feature. The marker does not say how small the edit was, so one AI-assisted fix labels the whole photo.
How Meta decides to add the label
Meta's labeling policy lists three sources:
- Industry-standard indicators in the file: C2PA and IPTC metadata. This is the most common cause on real photos.
- Your own disclosure: the AI label toggle in the post's advanced settings.
- Meta's own detection for content that has no metadata.
Since September 2024, posts that were only edited with AI show "AI info" in the post menu rather than as a visible tag, while content Meta believes is fully generated keeps the visible label.
The apps that most often cause it
Adobe Photoshop and Lightroom
Generative Fill, Generative Expand and Generative Remove can write Content Credentials and an IPTC value of compositeWithTrainedAlgorithmicMedia. Full details in the Photoshop and Lightroom guide.
Phone gallery AI edits
Google Photos Magic Editor, Samsung Galaxy AI generative edit and similar object erasers can tag the saved copy as AI-edited. A photo straight from the camera is fine; the edited copy is the one that gets flagged.
Canva, background removers and AI upscalers
Some of these write tool names or IPTC fields. Not all do, which is why checking the actual file beats guessing.
How to fix it, step by step
- Open the original file (not a screenshot) in the tool above and read the report.
- If you see an AI tag, press Clean & download.
- Post the clean copy as a new post. Cleaning cannot change a post that is already live; see removing AI info after posting.
- When posting, leave the AI label toggle off if no generative change shaped the image.
If the photo really was generated or meaningfully changed with AI, keep the label. The fix is for real photos that a blunt metadata flag got wrong.
When the report is clean but the label still appears
If the checker finds no AI markers, the label came from somewhere else: your own toggle, an invisible watermark, or Meta's classifier. Check the post's settings first. Otherwise, use Instagram's Help menu to report a problem with the label, and keep your original file as proof of what you shot.
Instagram AI info on real photos: FAQ
Why did Instagram label my photo AI info when I did not use AI?
Usually an app you used wrote an AI marker into the file. Generative erase, fill, expand, sky replacement and some background removers write C2PA or an IPTC source type, and Meta reads those signals when you upload. Run the photo through the checker on this page to see which field is responsible.
Does Instagram scan the picture itself to decide it is AI?
Meta says it uses industry-standard indicators in the file (C2PA and IPTC), people's own disclosures, and its own detection. For most mislabeled real photos, the metadata is the cause, which is why the label can appear on an image that looks completely ordinary.
Can a filter or a crop cause the AI info label?
A normal crop, exposure change or Instagram filter does not. Generative features do: anything that fills, removes or extends parts of the frame with a model can write an AI marker, even if the change is tiny.
Will removing the metadata stop the label?
If metadata was the trigger, a clean copy is usually not labeled when posted as a new post. It does nothing for a post that is already live, and it cannot beat invisible watermarks or Meta's classifiers.
Does the AI info label hurt reach?
Meta has not said the label reduces distribution. The bigger cost is trust: followers and clients may assume the photo is fake, which is why photographers want wrong labels gone.
Can my phone camera add the AI label?
A plain photo should not. AI edits made in the gallery app, such as generative erase or Magic Editor, can add an AI marker to the saved copy. Shooting is fine; check the file after editing.