Platform comparison
How Instagram, TikTok, LinkedIn and Pinterest Label AI Images
Each platform decides for itself when to call an image AI-generated. Some read C2PA Content Credentials, some read the IPTC digital source type, some run their own classifiers, and all of them rely on creators to disclose. This page compares what each one reads, and links to a guide for each.
Reviewed
What each platform reads
Metadata signals are the ones a cleaner can change. Watermarks, classifiers and your own disclosure are not.
| Platform | Label shown | Reads C2PA | Reads IPTC | Own detection | Keeps metadata after upload |
|---|---|---|---|---|---|
| Instagram “AI info” | AI info | Sometimes | Yes | Own watermark on Meta AI images; self-disclosure | No |
| TikTok “AI-generated” | AI-generated | Yes | Not documented | Own watermark (testing); creator labels | Yes |
| LinkedIn “CR” icon | CR icon and details panel | Yes | Not documented | None documented | Sometimes |
| Pinterest “AI modified” | AI modified | Sometimes | Yes | Classifiers | Not documented |
“Sometimes” marks cases where the platform's statement and independent tests disagree, or where behavior depends on the upload path. Reviewed September 25, 2026.
Instagram “AI info” Label: Why It Appears and What You Can Control
Instagram and Facebook add “AI info” when a file carries IPTC or C2PA AI signals, even after a small AI edit. See what triggers it and check your file first.
Read and cleanTikTok “AI-generated” Label: What TikTok Reads and What You Must Disclose
TikTok auto-labels uploads that carry C2PA Content Credentials and is adding invisible watermarks to AI content. See what it reads and what creators must disclose.
Read and cleanLinkedIn's “CR” Content Credentials Icon: What It Shows
LinkedIn shows a CR icon on images and videos with C2PA Content Credentials, and anyone can open the details. See what your file reveals before you post it.
Read and cleanPinterest “AI modified” Label: How It Works
Pinterest labels Pins “AI modified” using IPTC metadata, its own classifiers and creator disclosure. See what triggers the label and what file metadata changes.
Read and cleanThree kinds of signal
Metadata signals are the IPTC digital source type and C2PA manifests written by the tool that made or edited the file. They are easy to read and easy to lose: any re-encode drops them. Watermarks sit in the pixels and survive screenshots and compression. Classifiers judge an image by how it looks. On top of these, every major platform asks creators to label realistic AI content themselves.
That is why removing metadata can change an automatic label caused by an editing tool, but does not change a platform's rules. If your content is AI-generated and the platform asks you to disclose it, you still need to.
What changes in 2027
From 1 January 2027, California requires large online platforms to detect and display standards-compliant provenance data on uploads and not to knowingly strip it. Expect more platforms to show Content Credentials the way LinkedIn does, and to note when a file that should carry them does not.
The privacy side of uploads
Most large platforms strip EXIF and GPS from the copy the public sees, but they still receive the original file. Messages, document uploads and marketplace listings often keep metadata intact. Cleaning a file before you upload it is the only way to be sure your location and device details stay with you.
What a metadata cleaner can and cannot change
| Layer | What it holds | BlankAI |
|---|---|---|
| EXIF and GPS | Camera or phone model, serial numbers, capture time and location coordinates. | Removed |
| XMP and IPTC | Creator tool, edit history, captions, credits and the IPTC Digital Source Type AI label. | Removed |
| C2PA Content Credentials | Who or what created the file and how it was edited, signed by the issuing app. | Removed |
| PNG text chunks | Stable Diffusion prompts, seeds and model names, ComfyUI workflows. | Removed |
| Invisible watermarks | SynthID (Google and OpenAI images), Meta's Content Seal, TikTok and ByteDance watermarks. | Not removed |
| Visible watermarks | Logos, sparkle icons and corner badges drawn onto the image. | Not removed |
| AI classifiers | Models that judge an image by how it looks. File metadata plays no part. | Not affected |
Questions
Which platforms show Content Credentials to viewers?
LinkedIn shows a CR icon and a details panel on images and videos with C2PA credentials. TikTok and Meta use C2PA to apply their own AI labels rather than showing the credential itself.
Does removing metadata stop AI labels?
It removes the metadata signals only. Platforms also use watermarks, classifiers and your own disclosure, and their rules on labeling realistic AI content still apply.
Do platforms remove my photo's location?
Instagram, Facebook and public LinkedIn posts do not show EXIF or GPS to other users, according to independent tests, but the platform still receives the original upload. Clean the file first if you do not want to hand over that data at all.