Content Provenance (C2PA)
An open technical standard for attaching signed, tamper-evident provenance records to images, video, audio and documents.
Provenance records help brands prove their visuals are authentic, disclose AI-generated assets, and prepare for EU AI Act transparency duties that require synthetic content to be marked in a machine-readable way.
A marketing team generates campaign images with an AI tool that writes C2PA manifests, keeps credentials intact through the export settings of its editing software, and checks each file in a Content Credentials verifier before publishing.
Many platforms strip metadata on upload, and a missing manifest proves nothing about origin. C2PA shows who signed a claim, not whether the content is true, and watermark-based soft bindings can be removed or degraded.
A C2PA manifest, branded as Content Credentials, holds assertions about the asset, such as capture device, edit actions or AI generation, a claim that bundles them, and a claim signature made with an X.509 certificate. Hard bindings use cryptographic hashes to detect any change to the asset; soft bindings such as watermarks or fingerprints help recover credentials after metadata loss. Article 50 of the EU AI Act requires machine-readable marking of AI-generated outputs, a duty C2PA can help meet.
Cybersecurity, Experience Design
C2PA — C2PA Technical Specification 2.1 — https://spec.c2pa.org/specifications/specifications/2.1/specs/C2PA_Specification.html; EU Artificial Intelligence Act (Future of Life Institute) — Article 50: Transparency Obligations — https://artificialintelligenceact.eu/article/50/
