Why does my photo say “AI info”?
The seven most common reasons a normal-looking photo gets an AI label, from Generative Fill to phone eraser tools, and how to confirm which one applies to your file.
A label on a photo you took yourself feels like an accusation. It is usually something more boring: one editing step wrote a few bytes into the file, and the platform read them. Here are the causes we see most, roughly in order of how common they are.
1. Generative Fill or Generative Expand
Photoshop’s generative features attach Content Credentials and an IPTC source type saying part of the image was produced by a trained AI model. Extending a sky by a few centimetres is enough. This single cause was behind most of the “Made with AI” complaints in 2024.
2. A phone’s AI eraser or editor
Google Photos’ Magic Editor, Samsung’s Galaxy AI editing and similar tools on other phones record that AI editing was used. Removing a passer-by from a holiday photo can be all it takes. Samsung also adds a visible mark in the corner.
3. Background removal or replacement
Design and e-commerce apps that swap a background for a generated one mark the result. Plain cut-outs on white usually do not, but generated scenes do.
4. The image was AI-generated before you got it
Stock libraries now carry a lot of AI imagery, and templates in design apps often include it. If one element of your design came from a generator, its marker can travel into your export.
5. An “enhance” or upscaling tool
Some AI upscalers and portrait enhancers rebuild detail with generative models and tag the output. Others do not. It varies by app and version.
6. The AI label was switched on at upload
It is easy to tap by accident, and scheduling tools sometimes carry the setting from one post to the next. Tap the label on the post: platforms usually say whether it came from the creator.
7. The platform’s own detection
Pinterest in particular uses visual detection alongside metadata, so a very polished photo can be labelled with no marker in the file. Other platforms are adding invisible watermark checks as well.
How to tell which one it is
- Find the exact file you uploaded, not a screenshot of the post.
- Drop it into the AI label checker.
- Read the evidence list. A tool name tells you the app. “compositeWithTrainedAlgorithmicMedia” means a generative edit on real media. “trainedAlgorithmicMedia” means the file claims to be fully AI-generated.
- If nothing is found, look at causes 6 and 7.
Found the cause? The next step is in Real photo labelled as AI? Here is the proper fix. Adobe users should also read which Photoshop and Lightroom edits trigger a label.
What does not cause a label
Cropping, straightening, exposure, white balance, contrast, saturation, sharpening, noise reduction, lens corrections, standard presets and filters, text overlays and collages of real photos. Phone features such as HDR, portrait blur and night mode are computational photography, not generative AI, and are not treated as AI content.
Questions people ask
I never used AI. Why is my photo labelled?
You may have used it without thinking of it as AI. Object erasers, background replacement, generative fill and "expand" features all count, and they record their use in the file.
Does cropping or adjusting brightness trigger an AI label?
No. Ordinary adjustments such as crop, exposure, colour, sharpening and standard filters do not write AI markers.
Does Photoshop always trigger the label?
No. Only generative features do. Classic tools such as Clone Stamp, Healing Brush, curves and layers leave no AI marker.
How can I see what is in my file?
Use the AI label checker on this site. It lists every AI-related marker and where it sits in the file.