TagLens

Guides, updated 21 November 2026

Photoshop Generative Fill and the “AI info” label

Photoshop and Lightroom are the most common reason a real photo ends up with an AI label. When you use a generative feature, Adobe records it in the file’s Content Credentials, and platforms like Instagram, Facebook, TikTok and LinkedIn read that record.

Which features add the declaration

Classic tools like the Healing Brush, Content-Aware Fill (the non-generative one), adjustments, masks and crops don’t declare AI.

What the file actually says

A photo with one generative edit is usually marked as compositeWithTrainedAlgorithmicMedia, meaning a real capture with a generated part. A fully generated Firefly image is marked trainedAlgorithmicMedia. The first gets Meta’s softer “AI info” in the menu; the second gets a visible label. More on how Meta treats each.

How to check

  1. Drop the exported file into TagLens.
  2. Look at Content Credentials → History. You’ll see each recorded step and the tool that made it, for example “Edited with Adobe Photoshop — Edited with generative AI”.
  3. The platform table shows what Instagram, TikTok, YouTube, LinkedIn and X are likely to do with it.

If the label doesn’t fit what you did

If you used a generative feature, the label is accurate, even for a small edit. If you want an export without it, redo that change with a non-generative tool (Healing Brush, Clone Stamp, classic Content-Aware Fill) and export again. Adobe adds credentials on its own when generative AI is used, so the honest fix is in the edit, not in stripping the file afterwards.

Check your own file. See which field triggers the label. Nothing is uploaded.

Check a file

Written by . Platform rules change often; if something here is out of date, open an issue.