Admissibility & QA

The admissibility gate sits between the Evidence Ecosystem and the Constitutional Stack. Evidence that is normalised and hash-verified is not automatically execution-eligible. The gate applies independent, fail-closed checks before any artefact may be used by a reasoning operation.

There is no override. The gate fails closed.

The three gate components

Normalised evidence
     │
     ▼
[Plausibility Matrix] P03 — GB2519799.7
Registry-verified claim-type score + Merkle proof
Below threshold → guaranteed non-execution
     │
     ▼
[Probability Certification] P04 — GB2519801.1
Tail-risk bound derived and cryptographically committed
Failure probability ≤ ε at confidence C
     │
     ▼
[Quantum Alignment] P02 — GB2519798.9
Quantum + classical scoring in parallel (classical fallback always active)
Committed only if equivalence confirmed — bit-identical across nodes
     │
     ▼
Execution-eligible artefact

Plausibility Matrix — P03 / GB2519799.7

Plausibility Matrix →

For every claim-type, a PlausibilityScore is retrieved from a versioned safety registry and verified with a Merkle proof of inclusion. Three zones:

Zone	Score	Outcome
Below lower threshold	< 0.3	Guaranteed non-execution — claim enters dissent archive with full audit record
Intermediate band	0.3 – 0.7	Mandatory human arbitration before execution
Above threshold	> 0.7	Proceeds to Probability Certification

Claims that fail enter the dissent archive — not discarded. When the registry is updated by authorised authorities, previously non-executable claims may become executable. This is the mechanical pathway for recall.

Probability Certification Engine — P04 / GB2519801.1

Probability Engine →

For each DivergenceGraph node, a certified tail-risk bound is derived:

Failure probability ≤ ε at confidence C

A cryptographic commitment is computed over the canonical serialisation. The deterministic arbitration state machine evaluates the bound. Failure, validation error, or non-identical recomputation: non-executable state. No override.

This is the mechanism that makes AIEP compliance machine-verifiable rather than policy-asserted. A regulator can recompute the bound from the committed artefacts and confirm the certification holds.

Quantum Alignment Layer — P02 / GB2519798.9

Quantum Alignment →

Where quantum hardware is available, the Quantum Alignment Layer runs quantum and classical scoring in parallel, canonicalises both results, computes a deterministic deviation metric, and commits the result only if equivalence is confirmed.

Classical simulation is always the fallback. The committed result is bit-identical across all distributed nodes regardless of quantum hardware variance.

This provides a future-proof scoring substrate: quantum advantage can be exploited when available, but every deployment produces the same verifiable committed result regardless of hardware.

Fail-closed by design

Condition	What happens
Plausibility score below lower threshold	Non-execution. Claim committed to dissent archive.
Plausibility score in intermediate band	Human arbitration required before any execution.
Probability certification fails	Non-execution. No override.
Non-identical recomputation by any node	Non-executable state. Both nodes flagged.
Quantum/classical deviation exceeds threshold	Result not committed. Re-evaluation required.

The gate is the enforcement mechanism for AIEP’s core promise: that every executed claim has passed independent, cryptographically verifiable admission control.

Evidence Layer — normalisation and evidence assembly that feeds the gate
Architecture — where Layer 3b fits in the seven-layer stack
Certification & Registry — how CertificateHash values prove gate passage
GENOME & Swarm — the frozen kernel that enforces gate invariants