Public-sector pilot audit — 30-item measurement scope

2026-08-09 · CSOAI — measurement, not claim

Scope: a 30-item measurement battery for Cabinet Office/GDS-lens AI deployments, sampled from the published instruments — the same harnesses behind the public benchmark. Measured instrument performance today: care-gate recall 100%, over-block 0%. Your stack gets the same treatment, measured — never a paper opinion.

The 30-item battery

Care floor public (12) — refusal on unlawful practices with public-service framing
EU AI Act tiering (6) — public-sector use case classification
ProvBench survival (5) — records integrity through real transforms
GSPC S-axis (4) — security posture
Two-sided refusal (3) — no over-block on legitimate public requests

Deliverables

Signed measurement report (Ed25519, recompute-able) for each of the 30 checks.
Two-sided refusal profile: harm caught, availability preserved.
A scorecard entry you can publish, or keep private.

Why this is not a certification

We measure; regulators and accredited bodies decide. Every figure is a measurement on an open harness, signed and independently re-runnable — the same discipline as the public benchmark. ProvBench exists because the provenance of the evidence itself must survive; it does not here either, until it is measured.