Content provenance

Content provenance is cryptographic metadata attached to media that records how it was created, by whom or what model, and what edits it has been through.

Provenance in 2026 is dominated by C2PA (Coalition for Content Provenance and Authenticity), a manifest format that signs the chain of custody of an asset. Adobe, Microsoft, OpenAI, Leica, Sony, BBC are all C2PA members; major image generators (DALL·E 3, GPT-Image, Adobe Firefly) emit C2PA manifests by default. The chain answers: was this asset originally captured by a camera or generated by AI; which model; which edits applied; was it re-encoded. Provenance is a complement to watermarking — it lives in metadata rather than in pixels and survives differently. Both are needed to credibly track AI-generated media in 2026.

When to use content provenance

• Newsroom workflows.
• Stock image / commercial-license pipelines.
• Legal evidence chains involving images / video.

Common mistakes

• Treating provenance as proof of authenticity — it proves history, not truth.
• Stripping metadata on re-encode and losing the chain.

FAQ

Related terms

• AI watermarking — AI watermarking embeds invisible-to-humans signals in model output (text, image, audio, video) so the content can later be detected as AI-generated.
• Multimodal model — A multimodal model accepts more than one input type — text plus images, audio, or video — and reasons across them in a single forward pass.
• Guardrails — Guardrails are deterministic checks layered around a language model to prevent unsafe, off-topic, or non-compliant outputs from reaching the user.

Sources

• C2PA specification

Last updated: 2026-06-01. Raw markdown: https://promtable.com/glossary/provenance.md.