C2PA Content Credentials
C2PA Content Credentials are tamper-evident metadata attached to a media file that record how it was made, including whether AI was involved.
C2PA Content Credentials are a set of tamper-evident metadata, defined by the Coalition for Content Provenance and Authenticity, attached to an image, video or audio file to record its origin: which tool or camera created it, what edits were made, and whether generative AI was involved at any step. If the file is altered after signing, the credential shows the mismatch instead of silently disappearing.
The coalition backing the standard includes Adobe, Google, Microsoft, Meta, OpenAI and TikTok, alongside camera makers such as Sony, Nikon, Canon and Leica that now sign photos with Content Credentials at the moment of capture, and a growing list of newsrooms that sign published articles the same way. By 2026 the specification has passed version 2.4 and the coalition counts several thousand member and affiliate organizations.
Content Credentials are the technical mechanism many platforms rely on to detect and label AI-generated media automatically, even when a creator does not disclose it themselves, since the credential is designed to survive normal re-encoding and re-uploading.
Examples
- A news photo shot on a Leica M11 carrying a signed Content Credential from the moment the shutter clicks, before any editing happens.
- An AI-generated video keeping its C2PA marker intact after being re-encoded for 3 different platforms during publishing.
Related terms
AI-generated content labeling means marking media that was created or meaningfully altered by AI, both visibly on the post and invisibly in its metadata.
An AI agent is software that plans and carries out a multi-step task on its own, choosing and calling tools, instead of just answering one prompt.