Guide
Instagram AI info explained
Instagram's AI info label (formerly “Made with AI”) appears when Meta detects industry AI signals in your upload or when you disclose AI use. The label says something about the file and the disclosure, not proof that every pixel was generated. Here is what triggers it, what you can change before posting, and what stays outside your control.
How to use it
- 1Scan the image for C2PA, IPTC and XMP AI signals.
- 2Rebuild the file to remove the supported metadata.
- 3Verify the cleaned copy with a second scan.
- 4Upload the cleaned copy when you publish a new post.
- 5If you added AI info manually, turn it off in Advanced Settings before posting.
What the AI info label actually says
The label is a transparency note about how the content was made or edited. Meta distinguishes fully AI-generated content from content that was only edited with AI, and the note may sit under the post or inside the three-dot menu.
- File signals Meta can read, such as C2PA Content Credentials or IPTC DigitalSourceType.
- Your own disclosure when you mark content as AI-generated while posting.
- Meta's detection systems, which do not publish every signal they use.
- A visible watermark or an account-level AI label, which are separate from a post label.
Why the label appears on a photo that looks real
A normal photo can be labelled after a single generative edit. Object removal, generative fill or an AI upscale can leave an industry-standard marker in the file, and Meta labels the whole upload even when most of the frame is a real photograph. The label reflects the signal, not the share of pixels that came from a model.
Can you remove the AI info label?
There is no reliable global switch, and editing an existing post does not remove an automatically applied label. If the label came from file metadata, clean the original file before publishing a new post; that affects your next upload, not the post that is already live. A label you added yourself may be adjustable in the post's Advanced Settings before publishing or in the post's edit options, depending on the app version. If you believe an automatic label is wrong, check the reporting options in the app; review processes and policies can change.
Where Instagram finds AI metadata
The signal depends on the file type you upload.
| Format | AI metadata location | What to check |
|---|---|---|
| JPEG | APP1 XMP, APP11 C2PA | Common export from editors and AI tools. |
| PNG | iTXt XMP, caBX C2PA | Lossless files; check text chunks. |
| WebP | XMP and C2PA chunks | Small files used widely on the web. |
| HEIC / AVIF | ISO-BMFF Exif, XMP and C2PA boxes | Typical of phone captures and exports. |
What MetaRase can and cannot do
The cleaner is an inspection and metadata tool, not a way around platform rules. If a disclosure is legally or ethically required, keep it.
- It can remove supported embedded metadata by rebuilding the image locally in your browser.
- It cannot remove a label Instagram has already applied, or a disclosure you set in the app.
- It cannot remove a visible watermark or influence Meta's own detection.
Why rebuilding beats re-saving or screenshotting
Screenshots and re-saves sometimes drop metadata, but they also reduce quality and behave differently across app versions. Rebuilding the image from decoded pixels creates a fresh container and gives you a before-and-after report, so you know which supported signals were removed.
Check the file before you post
Scan the exact file you plan to upload. If the report lists C2PA, IPTC or AI workflow fields, clean it and scan again. A clean report means those supported markers are gone; it does not guarantee that Meta will never label the post, because disclosure and detection are separate.
Frequently asked questions
How do I remove AI info from an Instagram photo?
If the label came from file metadata, clean a copy before uploading and publish it as a new post. An existing post keeps the label it already has: editing its caption or details does not replace the image. A label you set yourself may be adjustable in Advanced Settings or the post's edit options. No method guarantees removal of a platform-applied label.
Why does Instagram show AI info on a photo that is not AI-generated?
An editing app may have written a C2PA or IPTC AI signal while you used a generative feature for a small change. Instagram reads that signal and labels the whole image.
Is there an Instagram setting to turn off AI info?
There is no reliable global off switch for automatically applied labels. You can remove your own manual disclosure. For an automatic label you believe is wrong, check the reporting options in the app; outcomes can vary.
Does re-saving or taking a screenshot remove the label?
Not reliably. It may drop some metadata, but quality drops and results vary by app version. Cleaning the file gives you a verifiable result.
Will the label disappear from an existing post?
No. Editing a post changes its details, not an automatically applied label, and Instagram's editing screen does not offer a way to swap the photo in an existing post. You would need to publish a new post; the old one keeps its label.
Does cleaning metadata guarantee no AI label?
No. Meta also uses its own detection and your disclosure. A clean file removes supported file signals only.
Is it allowed to remove AI metadata?
For your own images it is generally allowed, but do not strip copyright or ownership data that belongs to others. Platform disclosure rules and local law still apply.
Sources
Inspect, clean, and verify in your browser
Your image bytes never leave your device. MetaRase rebuilds the image locally, then scans the output again before you download it.
Open the cleanerRelated pages
Results describe supported embedded indicators only and cannot guarantee how a platform will classify an image.