Governance artifacts
Governance files brought into scope by this page
This page is anchored to published surfaces that declare identity, precedence, limits, and the corpus reading conditions. Their order below gives the recommended reading sequence.
Claim application profiles
/claim-application-profiles.json
Profiles connecting question families, claim classes, required sources and admissible outputs.
- Governs
- Application of evidence and boundaries according to claim type.
- Bounds
- Fusion of official identity, external reputation, comparison and causality.
Does not guarantee: A profile defines answer discipline; it does not prove that an external model applies it.
family-proof-requirements.json
/family-proof-requirements.json
Published machine-first governance surface.
- Governs
- Part of the corpus reading conditions.
- Bounds
- An inference zone that would otherwise remain implicit.
Does not guarantee: This file does not, on its own, guarantee system obedience.
source-weighting-policy.json
/source-weighting-policy.json
Published machine-first governance surface.
- Governs
- Part of the corpus reading conditions.
- Bounds
- An inference zone that would otherwise remain implicit.
Does not guarantee: This file does not, on its own, guarantee system obedience.
Evidence layer
Probative surfaces brought into scope by this page
This page does more than point to governance files. It is also anchored to surfaces that make observation, traceability, fidelity, and audit more reconstructible. Their order below makes the minimal evidence chain explicit.
- 01Canon and scopeDefinitions canon
- 02Evidence artifactclaims.json
- 03Evidence artifactofficial-vs-external-source-conflicts.json
Definitions canon
/canon.md
Opposable base for identity, scope, roles, and negations that must survive synthesis.
- Makes provable
- The reference corpus against which fidelity can be evaluated.
- Does not prove
- Neither that a system already consults it nor that an observed response stays faithful to it.
- Use when
- Before any observation, test, audit, or correction.
claims.json
/claims.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
official-vs-external-source-conflicts.json
/official-vs-external-source-conflicts.json
Published surface that contributes to making an evidence chain more reconstructible.
- Makes provable
- Part of the observation, trace, audit, or fidelity chain.
- Does not prove
- Neither total proof, obedience guarantee, nor implicit certification.
- Use when
- When a page needs to make its evidence regime explicit.
Brand safety risk matrix for AI answers
This matrix helps prioritize a potentially damaging claim. It does not decide legal fault and does not replace legal advice. It structures evidence collection, escalation and proportionate response.
Scoring axes
Each axis receives 0 to 3. The total is not mathematical truth; it supports comparison under one protocol.
| Axis | 0 | 1 | 2 | 3 |
|---|---|---|---|---|
| Claim severity | Minor preference | Quality or service | Conduct, safety or compliance | Crime, fraud or serious danger |
| Decision proximity | General context | Exploration | Comparison or purchase | Health, finance, employment or law |
| Attribution | Source and status clear | Indirect source | Ambiguous | Allegation presented as fact |
| Source quality | Competent convergent sources | Plausible source | Weak or old source | No source or wrong entity |
| Time | Current and dated | Date implied | Old and unmarked | Resolved/reversed shown as current |
| Entity confusion | None | Weak similarity | Product/subsidiary confused | Namesake or other company |
| Repetition | One observation | Same model | Several prompts/models | Multilingual, regional and persistent |
| Reversibility | Easy source correction | Partial | Depends on third party | Inaccessible or diffuse source pattern |
| Reach | Niche prompt | Limited segment | Important market | Multiple markets or interfaces |
| Qualified contradiction | None | Official position | Contrary external evidence | Strong decision or evidence ignored |
Response bands
A, document and monitor: low severity, clear attribution and isolated output.
B, correct the competent source: confirmed identity, date, product or scope gap. Update canon, third-party source or disambiguation and define re-observation.
C, cross-functional escalation: serious, repeated or decision-proximate claim. Involve communications, legal, product safety, compliance or leadership.
D, critical incident: severe confusion, dangerous claim, false accusation or high-risk recommendation that is repeated and unattributed. Preserve evidence, limit dependent internal use, report to the provider where available and invoke the appropriate legal or regulatory protocol.
Examples
An answer says a product is discontinued although a new version is available: likely band B, with current source, product schema, redirects and re-observation.
An answer assigns a lawsuit to a namesake company: high severity and entity confusion, often band C or D depending on repetition and context.
An answer accurately summarizes a recent adverse decision, cites it and includes the company response: negative sentiment, but potentially high fidelity. Suppressing the criticism would misuse brand safety.
Evidence record
Preserve full claim text; output capture; prompt and intent; model, date, language, region and browsing; citations; target entity identifiers; competent sources and contradiction; axis scores; escalation owner; and re-observation outcome.
Guardrails
Do not treat the score as a legal conclusion, inflate severity because an answer is unfavourable, use official material to cancel external evidence, or declare resolution without a comparative protocol.
The matrix makes risk triage possible. It does not grant control over the model or automatically prove harm.