Learn about AI labels & content credentials

On images, videos, and audio generated by Gemini in supported Workspace apps, you will see "Content Credentials", a technology that provides media history and transparency about the use of artificial intelligence (AI).

Understand AI labels & content credentials

Important: If the app is listed as "Google Media Processing Services", it means the file was definitely made or changed using AI. This is the official digital signature for content produced by Google's generative models.

Hidden digital labels are inside supported files, like images, videos, and audio. They give you details about:

Provider: The organization that created or published the content (e.g., Google).
App: The specific software or tool used to generate or modify the content.

Content credentials can:

Provide details about the origin and history of a photo or video through a standard that works across devices and apps from different companies.
Provide information from the app or device maker about AI or non-AI tools used to create or edit a photo or video.

Content Credentials cannot tell you if a photo or video is real or fake, but you can use the information together with the content of the photo or video to make informed decisions.

Inspect credentials within supported editors

For files opened within the Workspace apps listed below, you might find more specific content credential details. Use the entry point for your editor to view credentials:

Docs, Slides, Vids (Web): Right-click/context menu Content credentials sidebar.
Sheets (Web): Click the 3-dot menu (floating images) or context menu below "Put image over cells" (embedded images) Content credentials side panel.
Google Pics: Main menu Content credentials.
Drive Audio Overviews: open File Details to show the Details side panel.

Understand technical limitations

While content credentials offer insights into an asset's history, they operate within specific technical boundaries.

Platform & encryption limitations

Web-only inspection: Content credentials inspection is only available when using Google Workspace on a computer web browser.
No client-side encryption (CSE) support: Client-side encrypted files do not support C2PA signing.
No offline support: Offline files do not support C2PA signing.

Recognize fragile file history

C2PA metadata can be lost or altered. The "chain of custody" may be broken if:

You take a screenshot: A new file is created without the original credentials.
You resize or re-save: Using standard image or video editing tools can strip embedded metadata.
You use third-party tools: Sharing content via certain social media platforms often removes or modifies original metadata.
You download or export the file.

Tip: SynthID metadata is preserved during most edits, downloads, and exports.

Learn the difference between content credential history and facts

A valid content credential confirms how a digital file was made or edited. It does not confirm the content's truth or reality.

For instance, an AI tool might generate a realistic image of an event that never happened. While the content credentials will confirm the image was created by a specific AI tool, they do not verify the truthfulness of the image's subject matter.