Skip to main content
Glossary · Definition

Content provenance (C2PA)

Content provenance is a record of where a piece of digital media came from and how it was edited; C2PA is the open technical standard that attaches such a record, called Content Credentials, to images, audio and video.

Last updated October 3, 2026

How C2PA works

C2PA stands for the Coalition for Content Provenance and Authenticity, a group formed in 2021 whose early members included Adobe, Microsoft, the BBC, Intel and Arm. It publishes an open specification. A camera, editing tool or AI generator that supports it adds a signed manifest to the file, recording details such as the capture device or generating tool and later edits. The signature makes tampering with the record detectable. A viewer can then display the history, which the project brands as Content Credentials.

Provenance is the opposite approach to detection. A detector guesses whether content is synthetic. Provenance records how it was made, when the maker chooses to record it.

Limits

Support is uneven and changes over time, so check the C2PA site for current implementers. The record can be stripped when a file is re-saved, screenshotted or uploaded to a platform that removes metadata, and the absence of credentials proves nothing, because most media still has none. A valid credential shows what the signer claims about the file, not that its content is true. Related techniques such as invisible watermarks work differently and have their own limits.

Example

A hypothetical example: a communications team publishes executive photos with Content Credentials attached by its photographer's editing software. When a doctored version circulates, the team points reporters to the original with its signed history. The doctored file has no credentials and the original does, which supports the team's account but does not settle it for audiences who never check.

Why it matters to communications and risk teams

Provenance gives an organization one more way to show which media is its own, alongside the traditional practice of publishing originals on official channels. Teams verifying incoming media can check for credentials as one signal, next to tracing the earliest posting and contacting the supposed source. See deepfake, AI-generated content and fact-checking.

Common misconceptions

Content Credentials record how a file was made. They do not detect deepfakes, and they do not label content true or false.

Support is neither mandatory nor universal. The record only survives where tools preserve it, and it only helps where viewers look for it.

See how these signals show up in your own coverage on the PeakMetrics platform or run the free AI Perceptions check. Back to the glossary.

Track these signals in your own coverage.

PeakMetrics follows news, social and broadcast sources, and what AI assistants say about your organization.