Photoshop Generative Fill and the AI label.

What Adobe attaches, and why you can't switch it off

Adobe's help center says: "When qualifying generative AI tools are used to create or modify content in Adobe apps, C2PA metadata will be automatically attached to the file" (Adobe, updated August 2026). Asked whether it can be disabled, Adobe answers: "No. Under the new laws, this metadata cannot be disabled for qualifying content", pointing to the California AI Transparency Act and the EU AI Act.

According to Adobe, the metadata records:

• The application used
• The generative AI model used
• Whether the content was wholly generated or modified using generative AI
• The date and time of generative AI use
• A unique content identifier

Two clarifications from the same page: this applies only to generative AI tools, "not traditional AI-powered features", and Adobe "does not automatically add visible AI labels to content". Platforms decide whether and how to display the information. Lightroom works the same way: Adobe announced that Content Credentials are automatically attached to photos edited with Generative Remove (Adobe, May 2024).

How Instagram and Facebook treat an edited photo

This is where a small retouch and a fully generated image part ways. Meta says content with signals that it was created with AI is labeled "AI content", while content with signals that it was modified using AI "will not have the AI content label directly on the post, story, reel or thread" (Meta Help Center). In other words, a photo where you used Generative Fill to remove a stray object is treated as modified, and the information goes into the post menu rather than onto the photo.

Other platforms have their own rules. YouTube says it may automatically label content containing C2PA metadata, and that such labels cannot be adjusted by the creator (YouTube Help). LinkedIn shows a C2PA icon on cryptographically signed images and videos, visible to all viewers (LinkedIn Help).

Check your export

1. Export from Photoshop or Lightroom as you normally would, then drop the exported JPG, PNG or WebP into the box above.
2. The report shows whether a C2PA manifest is present, whether it is marked AI-generated or AI-edited, and which tools it mentions, such as Adobe or Photoshop.
3. If you decide to publish without the manifest, download the cleaned copy. The image data is copied byte for byte, and your color profile is kept, so wide-gamut exports don't shift color.

Firefly and other Adobe apps

Adobe's statement covers "qualifying generative AI tools" used "in Adobe apps", not just Photoshop (Adobe), so images generated in Firefly carry the same kind of record. Because the metadata states whether content was wholly generated or only modified, a Firefly image and a lightly retouched photo can look very different in the report. Drop both kinds of file in and compare the AI source line.

Think before removing it

There are good reasons to keep Content Credentials. They are how photographers and newsrooms show what was and wasn't changed, and Photoshop also lets you enable Content Credentials for ordinary edits so that your authorship travels with the file. If you work for clients, agencies or publications, check their requirements before removing it.

Removing the manifest is reasonable when the generative edit was minor, you disclose AI use where it matters, and you would rather not publish an edit history with every copy, or when you want all metadata gone for privacy. For a deeper look at the manifest itself, see the C2PA page.