The problem they address

It has become hard to tell by eye whether an image was photographed, edited or generated by AI. Content Credentials are an attempt to answer a narrower, checkable question: what does the file's own signed history say about where it came from?

The technology behind them is an open standard developed by the Coalition for Content Provenance and Authenticity (C2PA). "Content Credentials" is the friendly name for a C2PA record attached to a file.

How they work

A C2PA-aware camera, app or AI generator attaches a manifest to the file. It can contain statements such as:

  • which device or software created the file,
  • which edits were made, for example cropping or colour adjustments,
  • whether generative AI was used,
  • earlier versions ("ingredients") that the file was built from.

The manifest is cryptographically signed and bound to the content. If someone changes the image without updating the record, verification shows that the content and its credentials no longer match. Viewing tools display the information, often behind a small "CR" icon or an "About this image" panel.

Who is using them

Adoption has grown during 2026. In May 2026, Google said it was expanding content transparency and verification across Search, Gemini, Chrome, Pixel and Cloud, including verification of C2PA credentials and its own SynthID watermarks, with some features rolling out over the following months. Several AI image providers and camera makers also attach C2PA data to the files they produce.

Regulation is pushing in the same direction. In the European Union, transparency obligations for AI-generated content under Article 50 of the AI Act apply from 2 August 2026, and the European Commission's Code of Practice on Transparency of AI-Generated Content describes practical ways to mark such content in a machine-readable form. Provenance metadata is one of the techniques discussed. See our news article on the EU AI Act content-labelling rules.

What credentials prove, and what they do not

  • They show the history the signer recorded. A valid credential means the record has not been altered since it was signed.
  • They do not prove that the content is true. A real photo can still be misleading; a credential describes origin, not accuracy.
  • Missing credentials prove nothing either way. Most images online have none, including most genuine photos.
  • They depend on trust in the signer, much like a website certificate depends on the issuer.

Why credentials often disappear

Content Credentials live in the file's metadata. Many everyday actions create a new file without them:

  • taking a screenshot,
  • uploading to services that strip metadata,
  • re-encoding an image in an editor or converter.

This includes Convertios tools that re-encode images. The image compressor does not keep original metadata when it re-encodes a picture, and the Image Converter creates new files. That is useful for removing location data, as explained in our guide to photo metadata and privacy, but it also means C2PA information from the original is not carried over. Convertios does not create, read or verify Content Credentials. If provenance matters for your work, keep the original file and use C2PA-aware tools for editing.

Watermarking systems such as SynthID take a different approach by embedding a signal in the pixels or audio itself, designed to survive some edits. Platforms increasingly combine both methods.

How to check an image

  1. Look for a "CR" icon or "About this image" option where the image is displayed.
  2. Use a C2PA verification tool and upload or link the original file.
  3. Read the manifest: who signed it, which tools were used, whether AI is mentioned.
  4. Treat the result as one piece of evidence alongside the source, context and other reporting.