Governed evidence admission
Aurora-Lens governs evidence admission, persistent PEF state, provenance, authority and freshness, contradiction preservation, unresolved references, text and voice communication admissibility, and lawful continuation.
A bounded, non-exclusive research collaboration examining the boundary between governed evidence admission and execution authorisation.
How should an AI system determine which evidence objects, interpretations and communications may enter governed persistent state and become candidates for consequence?
Aurora-Lens governs evidence admission, persistent PEF state, provenance, authority and freshness, contradiction preservation, unresolved references, text and voice communication admissibility, and lawful continuation.
VALO adds authenticated identity, mandate and delegation, execution context, consequence and operational state before REHT makes the separate final authorisation decision and Veritas preserves verifiable receipts.
The collaboration defines a directional relationship without combining ownership or granting one architecture authority over the other.
A rejected evidence object, interpretation or candidate communication cannot be admitted by VALO, REHT or an execution component.
Admission only creates a candidate for further consideration. REHT may still modify, defer, deny, step up or halt the resulting action or release.

Margaret Stokes is the independent architect and creator of Aurora-Lens. Her work focuses on governed evidence admission, persistent PEF state, provenance, contradiction preservation, communication admissibility and lawful continuation.