Photoshop generative features

At a glance

Images

Content CredentialsA signed record (C2PA) of how the file was made, sealed so any change shows.

For its apps in general, Adobe says that when qualifying generative AI tools are used to create or modify content, C2PA metadata is automatically attached to the file without changing how it looks, and that it cannot be turned off. This applies to generative AI tools only, not to traditional AI-powered features.

Adobe does not list which tools or output types qualify, and the Photoshop help page read does not mention Content Credentials.

Generative AI content labeling FAQAdobe

Quoted passage Introduction, paragraph 3; answer to 'Which content will have C2PA metadata attached?'

When qualifying generative AI tools are used to create or modify content in Adobe apps, C2PA metadata will be automatically attached to the file. C2PA metadata is a widely adopted, open industry standard for sharing information about content's origin and editing history. It is embedded in the file and does not change the appearance of the content." "C2PA metadata is automatically attached to content when qualifying generative AI tools are used to create or modify content in Adobe apps.

Generative AI content labeling FAQAdobe

Quoted passage Q: Does this apply to all AI tools in Adobe apps?

No. This applies only to generative AI tools, not traditional AI-powered features.

Generative AI content labeling FAQAdobe

Quoted passage Q: Can I turn automatic C2PA metadata off?

No. Under the new laws, this metadata cannot be disabled for qualifying content.

Adobe Generative AI User GuidelinesAdobe

Quoted passage 6. Content Credentials

Adobe may attach or publish Content Credentials for content created or modified with generative AI features to let people know it was generated with or modified by AI. You must not remove, alter, or disable any Content Credentials.

Invisible watermarkA signal woven into the pixels, frames, audio or wording, meant to survive copying.

Not stated. None of Adobe's pages listed under Sources read says so.

Adobe describes invisible watermarking only for Content Credentials a user applies through Adobe Content Authenticity, not for the credentials it applies automatically to generative output.

AI label in metadataA plain metadata field (IPTC Digital Source Type) saying AI made the content. Anyone can add or remove it.

Not stated. None of Adobe's pages listed under Sources read says so.

No Adobe page read names an IPTC Digital Source Type term.

Visible markA mark you can see, such as a logo or a sparkle in a corner.

Adobe says it does not automatically add visible AI labels to content.

Adobe points users to its tools for adding visible watermarks themselves, and distinguishes visible watermarks from Content Credentials, which are machine-readable provenance metadata.

Generative AI content labeling FAQAdobe

Quoted passage Q: Does Adobe automatically add a visible label to my content, in addition to C2PA metadata?

No. Adobe does not automatically add visible AI labels to content.

Watermarks for Generative AI content FAQAdobe

Quoted passage Visible watermarks section (table of resources by content type and app)

Adobe offers resources to help you add watermarks to your online content.

Watermarks for Generative AI content FAQAdobe

Quoted passage Q: What’s the difference between watermarks and Content Credentials?

Visible watermarks are visible markings placed on content. Content Credentials are machine-readable provenance metadata that provide information about how content was created and edited.

Tool named in metadataThe tool or model written into ordinary file fields, such as Software or Creator.

Adobe says automatically attached C2PA metadata may include the application used, the generative AI model used, whether the content was wholly generated or modified with generative AI, the date and time of use, and a unique content identifier.

Generative AI content labeling FAQAdobe

Quoted passage Q: What does C2PA metadata include?

Automatically attached C2PA metadata may include details about: Application used Generative AI model used Whether the content was wholly generated with generative AI or modified using generative AI Date and time of generative AI use Unique content identifier

Limits and detection

Checking tool from the makerA way the maker offers to check a file for its traces.

Adobe says C2PA metadata attached to content can be viewed by uploading the file to its Inspect tool.

Generative AI content labeling FAQAdobe

Quoted passage Introduction, paragraph 4

To view the C2PA metadata attached to the content, you can upload or drag and drop it into Adobe’s Inspect tool.

What the maker says about limitsIn the maker's own words.

Adobe says third-party platforms decide whether and how to display metadata, so labels may look different from one platform to another.

Generative AI content labeling FAQAdobe

Quoted passage Q: Will my content show a label when it’s shared on another platform?

When content appears on third-party platforms, such as social media sites, search engines, and content platforms, those platforms determine whether and how to display metadata information to their users. As a result, labels may appear differently from one platform to another.