Skip to content
hideslop

What Instagram's AI info label means and why you got it

AI info is Meta's AI disclosure on Instagram, Facebook and Threads. What triggers it, how it differs from the old Made with AI label, and what it does to a post.

5 min read

On this page7 sections
  1. What the label looks like
  2. A short history
  3. What triggers the label
  4. Which apps cause it
  5. How to see what your file will trigger
  6. What to do about it
  7. FAQ

Short answer. "AI info" is the label Meta puts on posts it believes were made or edited with AI. It appears when you disclose AI use yourself, or when the file carries industry-standard AI metadata (C2PA Content Credentials or the IPTC "AI generated" field) that Instagram reads at upload. It is the same label on Facebook and Threads. The label on a published post cannot be switched off; a clean file posted again has no label.

What the label looks like

On a post, "AI info" sits next to the username, where the location tag normally goes. Tapping it opens a short explanation of why the post was labeled. For content that was only edited with AI, Meta moved the label into the post's menu in late 2024, after photographers complained that small retouches were being presented as fully generated images.

The label is informational. It does not hide the post, and Meta has not said it limits distribution. Creators widely report that labeled posts underperform, but that comes from experience, not from a published rule.

A short history

  • Early 2024. Meta announced it would label AI images on Facebook, Instagram and Threads using "industry standard indicators" written by generators, starting with its own Meta AI images, which carry a visible "Imagined with AI" mark and invisible markers.
  • May 2024. The "Made with AI" label went live. Within weeks photographers found real photos labeled because an editing tool had written Content Credentials.
  • July 2024. Meta renamed it "AI info" and started relying more on the signals inside the file than on guesses about the content.
  • Late 2024 onward. Labels for AI-edited (rather than AI-generated) content moved into the menu. The underlying trigger has not changed: metadata plus self-disclosure.

What triggers the label

1. Your own disclosure

Instagram's composer has an AI label toggle under advanced settings. Meta's rules say you must use it for photorealistic video or realistic audio that was digitally made or altered. Switch it on and the post is labeled, with no file check involved.

2. Metadata inside the file

Meta reads two things at upload:

  • C2PA Content Credentials. A signed manifest attached to the file that records which tool made or edited it. ChatGPT, Gemini, Adobe Firefly, Photoshop's generative tools, Microsoft Designer, Meta AI and Samsung's Galaxy AI all write one. What is C2PA? covers the format.
  • The IPTC digital source type. A plain metadata field that can say "trained algorithmic media" (AI generated) or "composite with trained algorithmic media" (partly AI). Midjourney, Google's Magic Editor and many generators write it in the image's XMP.

Instagram strips metadata from the copy it serves to viewers, but it reads the original first. That is why the label can appear on a post that looks like an ordinary photo.

3. Meta's own detection

Meta says it may label content its systems or reviewers identify as AI. In practice this is uncommon for everyday posts and mostly concerns realistic content around news and politics.

Which apps cause it

App or deviceWhat it writesLabeled on Instagram?
ChatGPT imagesC2PA manifest naming OpenAIYes
Gemini and Nano BananaC2PA and IPTC metadata, SynthID, a visible sparkleYes
Adobe FireflyC2PA manifestYes
Photoshop generative fill or expandC2PA manifest on exportYes, even on a camera photo
MidjourneyIPTC digital source type in XMPYes
Meta AIVisible mark, invisible watermark, IPTC metadataYes, as "Imagined with AI"
Samsung Galaxy AI editsC2PA manifest plus a visible markYes
Pixel camera (Pixel 10 and later)C2PA manifest recording a camera captureNo, unless AI editing was used
Stable Diffusion, ComfyUIPrompt text in PNG chunks, no standard AI fieldUsually not

The last row matters: a file with no standard indicator gets no automatic label, even when it is obviously AI. The label is about metadata, not about how the picture looks.

How to see what your file will trigger

Before posting, run the file through the free AI label checker. It reads the Content Credentials, the IPTC source type, any AI generator names in the metadata, and whether the file carries camera data. It runs in your browser and never uploads the file. If it shows an AI signal, Instagram will too.

What to do about it

If the label is right for the content, leave it. Meta's rules on realistic AI content are clear, and disclosure protects your account.

If the label is wrong for the content (a retouched product photo, an illustration, a small AI cleanup), you have two choices. Post the file as it is and accept the label, or remove the metadata and post a clean copy. The step-by-step is in how to remove the AI label from an Instagram post. The image remover removes the manifest and the AI tags and re-saves the image as an iPhone photo; the first file is free.

FAQ

Is AI info the same as Made with AI?

Yes. Meta renamed "Made with AI" to "AI info" in July 2024 and later moved the label for edited content into the post menu. The trigger is the same: metadata inside the file or your own disclosure.

Does the AI info label apply on Facebook and Threads?

Yes. Facebook, Instagram and Threads share Meta's labeling system and read the same metadata.

Why did a real photo get the AI info label?

Something in the editing chain wrote AI metadata. Photoshop's generative tools, Lightroom's AI features and phone editors such as Galaxy AI and Magic Editor all do it. The photo is real; the file says it was edited with AI.

Can I turn the label off on a post that is already published?

No. Meta offers no way to remove an automatic label from a live post. Remove the metadata from the file and post it again.

Does the label appear on Stories and Reels?

Yes. Stories, Reels, carousels and single images are all checked the same way, and a Reel's video file carries the same kind of metadata as an image.

Your next post deserves a clean file.

Your first removal is free. No account, no card.

Check a file for AI labels