“AI info” on Instagram is Meta’s label for content that was created or edited with AI. It appears when Meta’s systems detect industry-standard AI signals in the file, such as Content Credentials or IPTC metadata, or when the person posting discloses that they used AI. Tap it to see more detail. The same label is used on Facebook and Threads. A missing label doesn’t mean an image is real, because the label depends on signals that are often absent or stripped.
Where the label came from #
Meta started labeling AI images in 2024. The label was first called “Made with AI.” On July 1, 2024, Meta renamed it “AI info” after photographers complained that small edits were getting the same label as fully generated images (Meta).
On September 12, 2024, Meta changed where the label shows. For content its systems detect was only modified or edited by AI tools, the label moved into the post’s menu. For content detected as generated by AI, it stays visible on the post.
| Content | Where “AI info” appears |
|---|---|
| Detected as AI-generated | On the post itself |
| Detected as only edited with AI | In the post’s menu |
| Disclosed by the poster | Shown as self-disclosed |
How does Instagram know a post is AI? #
Meta relies on three kinds of signal.
1. Metadata from other companies’ tools #
Meta says it detects the “AI generated” information in the C2PA and IPTC technical standards, and it built tools to read those signals from Google, OpenAI, Microsoft, Adobe, Midjourney and Shutterstock as those companies add them to their images (Meta). If a file carries that metadata when it’s uploaded, Instagram can label it automatically.
2. Its own markers on Meta AI images #
Images made with Meta AI carry visible markers, invisible watermarks and embedded metadata, according to Meta.
3. What the poster says #
Meta asks people to disclose when they share realistic AI-generated or digitally altered content, and it has said it requires disclosure for photorealistic AI video and realistic-sounding audio because it couldn’t yet detect those signals at scale, with possible penalties for not disclosing. That was Meta’s position in 2024; check its current rules for video.
Why did my real photo get an AI info label? #
Usually because your editing software wrote an AI marker into the file. If you removed a stray object with a generative tool, extended a background, or used an AI enhance feature, the app may have recorded that in the metadata, and Instagram read it. The September 2024 change was meant to make these cases less prominent by moving the label into the menu.
To check what your file says before posting, read its metadata. How to check image metadata for AI shows how on a phone or computer. If you edited the photo in Google Photos with Magic Editor or Magic Eraser, Google marks those edits in IPTC metadata (Google), which is the kind of signal Meta reads. For more on false flags in general, see why real photos get flagged as AI.
What the label can’t tell you #
The label is only as good as the signals behind it. It misses a lot:
- Stripped metadata. Many generators don’t write AI metadata, and a screenshot or re-save removes what was there.
- Tools outside the standard. Images from tools that don’t add C2PA or IPTC markers won’t be detected by metadata.
- People who don’t disclose. Self-disclosure depends on honesty.
So an unlabeled image isn’t a verified one. Treat “AI info” as a helpful positive signal and its absence as no signal at all.
How to check an unlabeled Instagram post #
- Open the post full screen and look at the details: text, hands, reflections, lighting. AI image artifacts lists what to check.
- Look at the account. New, anonymous, and posting only dramatic images is a pattern.
- Reverse search the image to see if it’s an older photo or a known fake.
- Run an AI check. Screenshot the post or save the image, then share it to Expose AI. Crop away Instagram’s interface, tap Analyze, and you get Likely real, Uncertain or Likely AI-generated with a confidence meter. Screenshots lose the original metadata, so the app’s on-device model makes the call from the pixels, and the image stays on your phone.
For breaking-news images, follow the routine in how to check if a viral photo is real.
Frequently asked questions #
Does “AI info” mean the whole image is fake? #
No. It means Meta detected AI signals or the poster disclosed AI use. That can range from a fully generated image to a photo with a small generative edit. Tap the label to see what Meta says about that post, and remember that edit-only labels usually sit in the menu.
Does Facebook use the same label? #
Yes. Meta uses the “AI info” label across its apps, including Facebook, Instagram and Threads, based on the same detection and disclosure rules.
Can Instagram detect all AI images? #
No. Automatic labeling depends on metadata and watermarks that many images don’t carry, especially after screenshots or re-uploads. Meta itself relies on disclosure for content it can’t detect.
Is it against the rules to post AI images without a label? #
Meta asks users to disclose realistic AI-generated or altered content, particularly photorealistic video and realistic audio, and has said it may apply penalties when people don’t. Rules change, so check Meta’s current policies for what applies to your post.