The CAI wrote the vocabulary before the standard existed

The shot

On 15 April 2020, the Adobe-founded Content Authenticity Initiative held what its own blog describes as 'the first meeting towards the publication of a white paper outlining our approach to content attribution,' gathering, in the CAI's words, 'representatives from social media, software, media, academia and non-profit industries.' On 3 August 2020, the CAI published that paper, 'Setting the Standard for Content Attribution,' naming eleven authors from Adobe, the BBC, CBC/Radio-Canada, Microsoft, The New York Times Company, Stanford University's Center for Blockchain Research, Truepic and the University of California, Berkeley.

What the documents show

The CAI's blog post describes the paper as proposing 'a layer of secure provenance which expresses relevant facts about how media is altered from the moment of creation to the moment of audience experience,' built on two named constructs: assertions, which 'capture the who, what, and how of asset creation and modification,' and claims, which 'add a layer of cryptographic verifiability and trust.' A separate CAI FAQ page, retrieved 16 September 2026, states the CAI — founded by Adobe in 2019, which 'continues to lead it' — is distinct from the Coalition for Content Provenance and Authenticity (C2PA), the formally governed standards body that 'unifies the efforts of the Adobe-led CAI' with a separate Microsoft- and BBC-led initiative, Project Origin. The vocabulary the 2020 paper introduced is the CAI's own working terminology, not itself a citation to the later C2PA specification.

The workflow

The FAQ describes the CAI's role as driving 'feature explorations,' creating 'open-source tools,' and fostering 'community,' while the C2PA 'provides the end-to-end open technical standards.' In practice, a team building a Content Credentials feature typically starts with the CAI's open-source tools and community documentation, while a compliance review of what a specification formally requires belongs with the C2PA's own published spec. The FAQ names collaborators including Adobe, the Associated Press, the BBC, Microsoft, The New York Times Company, Reuters, Leica, Nikon, Canon, Truepic and Qualcomm — a membership list, not a conformance certification.

What the tool does not change

Editorially: participating in the CAI's discussion, or being named a collaborator, is not the same as a product being formally assessed against the C2PA's conformance program. The 2020 paper's own framing, that it represents 'the starting line of a long journey,' is a useful caution against treating a single white paper or working meeting as a finished standard.

• Is a company named as a CAI 'collaborator' also listed on the C2PA's own Conforming Products List, or only associated with the broader community?
• Does a tool's documentation trace its terms — 'assertions,' 'claims,' 'Content Credentials' — to the CAI's working vocabulary, the formal C2PA specification, or a vendor's paraphrase of both?
• When was the cited CAI resource last updated, given the initiative describes its own work as an ongoing, multi-year process?

The CAI's own record shows a community effort that supplied vocabulary and early tooling later formalized elsewhere. Citing the CAI as evidence of a standard being met confuses that community layer with the separate coalition that actually defines and tests conformance.