AI can simulate.
PlenaProof records who reviewed.
VRX-5 is the layer PlenaProof uses when AI or agents create ambiguity around authorship, review, consent, evidence, decision authority, or appealability.
VRX-5 protects against AI-era ambiguity.
VRX-5 adds AI-use disclosure, human-in-the-loop confirmation, synthetic media flags, model-output warnings, agent authorization records, provenance strength, and appeal/review fields.
Was AI involved?
Records whether AI generated, summarized, classified, recommended, translated, screened, or escalated any part of the workflow.
Who stayed responsible?
Records the human reviewer role, decision, evidence checked, override state, and appeal route.
What could be fake?
Flags synthetic media, model-generated outputs, low-provenance evidence, automated-agent action, and adversarial-risk concerns.
| VRX-5 field | Purpose | PLENA surface |
|---|---|---|
| ai_use_disclosure | Names where AI assisted the workflow. | AI Oversight Receipts; Individual AI Proof. |
| human_review_confirmation | Shows that a person reviewed before action. | Human Review Receipts; reviewer training. |
| synthetic_media_flags | Marks content that may be generated or manipulated. | DETECTA, PROVA, AI Harm Defense. |
| agent_authorization_record | Records what an AI agent was allowed to do. | AI Agent Receipts; Agent Action Registry. |
| provenance_strength_score | Signals how strong the proof chain appears. | Proof Packets, Public Verify, wallet timeline. |
PLENA routing
VRX-5 strengthens AI Oversight Receipts, AI Agent Receipts, Individual AI Proof, AI Harm Defense, Proof of Human Agency, DETECTA authenticity review, and sector AI trust suites.
Schema anchor
assets/vrx5-ai-trust-receipt.schema.json receipts/PLENA-DEMO-VRX5-20260516-001.json