Skip to content

Provenance as Theater

Public discussion often frames C2PA as a tool that will "prove authenticity" or "fight AI slop." The actual specification does neither of those things directly.

The standard lets someone attach a signed history to a file. Platforms can then read that history and decide what to do with it. A platform might display a badge, suppress content without a manifest, or route it into a different moderation queue. None of these outcomes are required by the specification. They are policy choices made by whoever controls the platform.

This creates a performance requirement. Artists who want their work to remain visible or credible on major surfaces may need to supply the correct manifest even when the manifest adds little information about the actual making of the work. The manifest becomes part of the delivery package rather than an independent record of authorship.

At the same time, the technical properties that make the manifest removable remain unchanged. A file with a clean manifest today can be stripped tomorrow and re-saved without it. The absence of a manifest is easy to achieve. The presence of one is easy to fake at the point of first publication.

The gap between the language used in announcements ("restore trust," "combat misinformation") and the operations the format actually supports is consistent. The format excels at creating an auditable chain once a claim has been attached. It does little to establish that the first claim was the true first claim, or that the signer is the person the claim describes.

For artists, the practical question is not whether the standard exists. It is whether attaching these claims improves their position relative to the systems that already devalue their labor. In many cases the answer is conditional and depends on how the platforms that adopt the standard choose to interpret its presence or absence.

CC BY 4.0 — Translations and contributions for other languages are welcome.