Minimum Boundary Review External Crosswalk

LUMINA-30 public HTML reference page.

Status

This document is a non-core, non-binding mapping reference for relating the LUMINA-30 Minimum Boundary Review to existing governance, audit, procurement, incident review, and risk-management workflows. It does not modify LUMINA-30 Core Terminology and does not create new binding obligations, legal authority, regulatory force, certification status, official adoption, institutional endorsement, or binding standard status.

In this document, "adoption" means optional reuse or mapping surface only. It does not mean institutional adoption, endorsement, certification, or approval of LUMINA-30.

This document does not replace NIST, ISO, OECD, EU AI Act, procurement, audit, or organizational risk-management processes. It identifies where a short LUMINA-30 boundary-responsibility question can be inserted into those existing processes.

Purpose

The least disruptive external reuse path is not to ask institutions to adopt the whole LUMINA-30 framework. The fastest path is to map one minimum boundary review question to review surfaces they already use.

Minimum insertion question:

Before irreversible consequences occur, does the process verify whether effective human refusal, correction, recovery, or rollback remains practically available, and whether responsibility for preserving those conditions is explicitly assigned?

External Crosswalk

The following table gives the English mapping first. A Japanese mirror table is provided immediately after it so that reviewers can use the same mapping without translating the core rows themselves.

Existing surface Insert the Minimum Boundary Review here What LUMINA-30 adds
NIST AI Risk Management Framework Govern, Map, Measure, or Manage review notes; risk register entries; incident learning records A boundary-responsibility check for whether effective refusal, correction, recovery, or rollback remains practically available before irreversibility
NIST Generative AI Profile Generative AI risk action tracking; deployment review; misuse and overreliance review A compact check for whether optimization-driven deployment removes human intervention conditions before irreversible effects
ISO/IEC 42001 AI management system Responsibility assignment, records, internal audit, corrective action, continual improvement A reviewable question about who preserves boundary responsibility, evidence, verification, and correction before irreversibility
OECD AI risks and incidents Incident and hazard evidence, reporting notes, accountability review, post-incident learning A post-incident boundary finding: whether effective refusal existed before irreversible consequences, and whether undefined responsibility contributed to failure
EU AI Act Article 27 fundamental-rights impact assessment Impact description, affected groups, risk measures, human oversight, mitigation review A pre-deployment question about whether affected humans retain practical refusal, correction, recovery, or rollback conditions before irreversibility
Internal AI governance review Project approval notes, model deployment review, high-impact system review A short boundary-responsibility field that can be added without replacing the review structure
Audit checklist Evidence review, responsibility review, corrective-action review A check for whether the audit verifies practical refusal and responsibility assignment before irreversible effects
Procurement review Supplier requirements, scope review, update controls, termination clauses A clause asking whether the supplier and deployer preserve refusal, rollback, evidence, verification, and correction responsibilities
Incident report Timeline, cause, impact, containment, lessons learned A boundary finding that distinguishes technical failure, responsibility diffusion, evidence absence, anti-bypass failure, and delayed correction
Organizational risk register Risk description, owner, controls, evidence, review date A minimal field for boundary risk, responsible actor, evidence, anti-bypass condition, and next review point

Japanese Mirror Table

Reference Surface Notes

These references are used as mapping surfaces, not as authorities that LUMINA-30 replaces. Update this crosswalk only when an official source changes materially or when a better stable official URL is confirmed.

Reference surface Current use in this document Japanese translation note
NIST AI RMF Existing AI risk-management surface AI
NIST Generative AI Profile Generative AI risk profile and action surface AI
ISO/IEC 42001 AI management system surface AI
OECD AI risks and incidents Incident, hazard, accountability, and reporting surface AI
EU AI Act Article 27 Fundamental-rights impact assessment surface

Minimal Field to Add

Use the following field when the existing process cannot accept a full LUMINA-30 form.

LUMINA-30 Minimum Boundary Review:
Before irreversible consequences occur, are effective human refusal, correction, recovery, or rollback practically available?
Responsible actor:
Evidence:
Anti-bypass condition:
Correction owner:
Decision tag: Preserved / Not Demonstrated / Not Applicable / Needs Review

How to Use This Crosswalk

  1. Identify the existing workflow already used by the institution.
  2. Select the closest row in the crosswalk.
  3. Insert the Minimum Boundary Review field at the corresponding review surface.
  4. Do not claim that LUMINA-30 replaces the existing framework.
  5. Record whether responsibility, evidence, anti-bypass review, and correction ownership are explicitly assigned before irreversibility.

Relationship to Other LUMINA-30 Materials

Non-Prescriptive Boundary

This crosswalk does not determine the correct legal, policy, technical, procurement, or enforcement mechanism. Those decisions belong to competent institutions in each domain.

LUMINA-30 only asks whether the boundary condition was preserved:

Were effective human refusal, correction, recovery, or rollback practically available before irreversible consequences, and was responsibility for preserving those conditions explicitly assigned?