Representational Integrity Domain Profile

Representational Integrity applies where AI materially creates, alters, infers, simulates, attributes, authenticates, distributes, or governs a representation linked to an identifiable person or authorized persona. Provenance evidence is appraised within a trust model; it is not automatic proof of truth or authority.

Verification requirements

ID	What must be established
`RP-01`	Authority is established for the subject, representation, purpose, context, audience, duration, transfer, and withdrawal or exception rules.
`RP-02`	Synthetic or materially manipulated representation is disclosed when source authority requires it or nondisclosure would create material deception.
`RP-03`	Provenance, credential, attestation, and authenticity claims identify their trust model and do not exceed the validated evidence.
`RP-04`	Identity, biometric, exemplar, credential, provenance, intent, and representational-use data remain necessary, bounded, protected, and controlled.
`RP-05`	The system does not materially misattribute speech, action, intent, endorsement, role, or authorization and preserves material context.
`RP-06`	Affected people and legitimate representatives can contest, correct, challenge provenance, request removal or restriction, and escalate.
`RP-07`	Safeguards address missing, stripped, broken, stale, forged, or conflicting provenance and foreseeable unauthorized reuse.
`RP-08`	Truth, authenticity, identity, consent, authorship, ownership, authorization, detection, completeness, and legal claims stay within appraised evidence and authority.

Cryptographic validity is evidence about a bounded claim. It does not replace appraisal of issuer, subject binding, status, context, policy, authority, or meaning.

Evidence and claim boundary

Evidence may include authority records, permissions, licenses, credentials, provenance manifests, attestations, audit logs, disclosure tests, watermark and detection evaluations, incidents, user research, and remedy outcomes.

W3C Verifiable Credentials, C2PA Content Credentials, RATS concepts, watermarking, perceptual matching, and other controls may be selected where fit; none is universally sufficient. This Profile does not itself prove truth, identity, consent, authorship, authenticity, ownership, legal effect, or legal compliance.

Composition and stewardship

Compose with Relational Architecture for persona and companion agents, Cognitive/Epistemic Coherence for provenance-dependent reliance, Developmental Interaction for minors, and applicable Source, Deployment, and Modality Profiles.

The Representational Integrity Division is the corresponding stewardship identity.