One-Question Pilot for AI Adoption Review

🛩️ Pilot Airfield
🧭 Return to Central Square

A low-risk way to test whether effective human refusal remains available before irreversible escalation.

A lightweight one-question pilotThis is not certification, approval, official compliance, a legal conclusion, a safety guarantee, or evidence of institutional adoption.

See other responses / leave an anonymous response

The one question

Add this question to one AI adoption review, procurement check, audit note, security review, or incident-review form:

Can a human still stop it before irreversibility?Before this AI-mediated process becomes irreversible, can an accountable human or oversight body still refuse, stop, suspend, review, or redirect it?

First, answer with three choices only: answer this one question as clear, unclear, or not currently supported. For this pilot, that is enough. Use the evidence fields only if internal explanation, audit, or procurement confirmation becomes necessary.

How to run the pilot

  1. Choose one AI system, vendor feature, agent workflow, automation layer, or incident scenario.
  2. Ask the one question before approval, deployment, expansion, or incident closure.
  3. Record the answer as clear, unclear, or not currently supported.
  4. If the answer is unclear, identify which part is missing: time window, authority, evidence, rollback path, appeal path, or accountable human owner.
  5. Do not claim LUMINA-30 compliance. Treat the result as a reference-use note.

The result is a one-line note you can drop directly into an existing review record — no new process or document required.

Optional: minimal evidence fields if detailed records are needed

This section is optional. In the first check, answering the one question in three choices is sufficient. Use the fields below only when the result needs to be retained for internal explanation, audit, procurement confirmation, or later review.

Open optional evidence fields
FieldWhat to record
System or workflowWhich AI-mediated process is being reviewed.
Irreversible stepWhat action, decision, deployment, credential change, production action, or rollback loss would become hard to reverse.
Human refusal pointWho can refuse, stop, suspend, review, or redirect before that step.
Time windowWhen the human refusal must occur.
EvidenceLogs, approvals, UI controls, runbooks, rollback plans, appeal routes, or review records.
ResultClear / unclear / not currently supported.

When to escalate

Escalate from this one-question pilot to a fuller boundary review if stopping authority, rollback, recovery, privileged actions, appeal, or evidence access is unclear.

Next if you read only one page

Adoption Review

Check how to add this one question to AI adoption, procurement, audit, and security review workflows.

Includes: where to add this question in vendor questionnaires, audit checklists, and security sign-offs.

Open Adoption Review

Use only if needed:

Public-Source Review Template

Use this when public information is the only available evidence.

Open Template

Incident Review Floor

Move from a one-question pilot to incident or boundary review.

Open Incident Review