Adobe recently released V1 of its own suite of generative AI tools called “Firefly” (waitlist only for now – I don’t have access yet, but managed to piece together this information). Designed to integrate with Adobe’s existing apps, Firefly enables users to create generative AI images. Adobe has also incorporated “Content Credentials” into its AI-generated image outputs, asserting that the content was produced by AI (see the screenshot above).

Content Credentials are Adobe’s in-house implementation of the Coalition for Content Provenance and Authenticity (C2PA) standard. The C2PA is an open industry initiative focused on developing standards for tracing the provenance and authenticity of web media. By incorporating Content Credentials into Firefly, Adobe aims to promote transparency and trust in the AI-generated media space.

As part of the Content Authenticity Initiative (CAI) metadata, Adobe has introduced a custom assertion: com.adobe.generative-ai. This assertion helps establish that the content was generated using Adobe’s Firefly AI tools.

A new addition to the C2PA standarddigitalSourceType: trainedAlgorithmicMedia—will further complement Adobe’s efforts in the AI-generated media space. This new classification highlights the role of trained algorithms in generating media content, further establishing transparency.

Although social media platforms have yet to widely adopt the Content Credentials system, end users can verify content credentials of files on Adobe’s website at to see how the records work. This allows users to understand the provenance of AI-generated media and promotes transparency. You can also use Content Credentials currently in Photoshop.

Adobe has taken measures to ensure that the training data for its AI-generated images is curated from Adobe Stock and other reliable sources. According to the Firefly site, the curation process aims to mitigate harmful and biased content while respecting artists’ ownership and intellectual property rights (read more on CNET).