Instagram's AI Info Label: Causes and Fixes
Instagram attaches an "AI info" label when it detects industry-standard AI indicators in what you upload — principally a C2PA Content Credentials manifest, XMP AI-edit tags, and Google's SynthID signal in the pixels. Meta applies it automatically. It does not ask, and it is reading the file rather than looking at the picture.
There are two separate problems here, and most guides only solve one. The first is the label sitting on a post right now. The second is the file itself, which will keep triggering the check on every future upload until the markers are gone. Below covers both, in that order.
Turning Off the Label on a Post That's Already Up
Instagram now exposes an edit path on some posts — try this before deleting anything.
Unlike TikTok, Instagram has added a way to remove the label from a published post. Open the post from your profile, tap Edit, select the AI label, and deactivate it from there.
This does not work everywhere. Availability varies by post type, account, and region, and Meta has changed the behaviour more than once during 2026 — so treat it as worth trying rather than guaranteed. If the option is not there, the alternative is deleting the post and re-uploading a cleaned file.
Removing the label from one post does nothing to the file on your computer. Upload the same image again and the check runs again, with the same result. Fixing the post and fixing the file are two different jobs.
Stopping It Happening on the Next Upload
Clean the file first — which layer you clear depends on what the file carries.
Instagram reads two independent layers, and they need different treatment.
The C2PA manifest and XMP tags sit in the file header. These are the common case, they come off easily, and clearing them resolves most AI-info labels. If your image came from Grok, Adobe, Firefly or a recent phone camera, this is probably all you have.
SynthID sits in the pixels. Google's image models write it, and ChatGPT images have carried it since OpenAI adopted it in 2026. It is built to survive re-saving, cropping and screenshots, so no metadata cleaner removes it. If you cleared the metadata and Instagram labelled the post anyway, this is why — and the fix is regenerating the image rather than editing it.
Check the file first so you know which you are dealing with, then clean the layer that is actually there.
Why Photographers Get Labelled Too
Routine AI-assisted retouching writes the same markers as full AI generation.
A large share of AI-info labels land on photographs nobody generated. Background Generative Fill on a portrait, an AI sky replacement on a landscape, Lightroom's AI Denoise on a high-ISO frame, a Reel cover cut in CapCut — each can write AI-edit markers into the exported file. Instagram reads the marker and labels the post, even though the image is a real photograph that was lightly retouched.
If that is your situation, the fix is usually the cheap one: the markers are header metadata, so clearing them is a fast pass on your regular allowance. You can check what your file carries first and confirm there is no pixel-level signal to worry about.
Reels, Carousels and Stories
Every file in a post is checked, including the cover frame — clean all of them.
Instagram checks each asset you upload, not just the first one. A carousel where nine images are clean and one still carries credentials can be labelled on the strength of that one. Reels are checked as video and the cover image is checked separately, so a clean clip with an AI-generated cover can still pick up the label.
For video, the practical layer is the container: content credentials and provenance metadata in the file, cleared without re-encoding so the picture and audio are untouched. Pixel-level SynthID in video is not something we remove yet, and we would rather say so than sell you a metadata wipe under a different name.
The habit worth building is simple: clean every asset in the post, verify each one, then upload.