# 09 — Operational Refusals and Dispute Panel

> *A narrow, specific, universal-harm list of operations the protocol refuses. A panel of multi-stakeholder reviewers handles disputes. Everything else is allowed.*

## The two mechanisms

The protocol ships with exactly two governance mechanisms for hostile or harmful content:

1. **Operational refusal list** — a small set of claim-shapes the protocol refuses to represent at the write layer. Narrow, specific, universal-harm, operation-based (not topic-based).

2. **Dispute panel** — a multi-stakeholder body that reviews edge cases, appeals against refusals or validator actions, and maintains the refusal list.

Everything outside these two mechanisms is permissionless: the chain accepts it, the federation displays it or doesn't according to aggregator policy, the consumer filters it according to CPML.

## Operational refusals — what they are, what they are not

The refusal list is **operational**, not **topical**. The distinction matters.

### Operational refusal (correct)

"The protocol refuses to represent an attestation of the form *'I verify that <URL> contains CSAM'*." Reason: the operation of verification requires possessing the referenced content; possession of CSAM is itself a violation. The operation is incoherent, and the protocol refuses incoherent operations.

Other operational refusals likely to be on the narrow list:

- Attestations verifying the authenticity of non-consensual intimate imagery.
- Attestations verifying detailed bio-weapon synthesis instructions (operation requires the instructions to exist in verifiable form, which the protocol does not want to make discoverable).
- Attestations on specific-individual-targeted threats (e.g., "this threat against X is credible" — the protocol is not a threat-assessment service).
- Attestations of the form "this illegal market is operating at <URL>" (operation is a referral to illegal content).

### Topical refusal (incorrect)

"The protocol refuses to represent claims about vaccines." Reason: vaccines are a legitimately-contested topic across scientific, medical, religious, political, and public-health frames. Refusing to represent claims about them imposes a specific frame as correct and the others as forbidden.

Similar topical questions the protocol does *not* refuse:

- Claims about climate science (contested under some frames; the protocol records both).
- Claims about election integrity (contested; recorded under multiple frames).
- Claims about historical atrocities (the claim may be true under mainstream frames and denied under revisionist ones; the protocol records both; mainstream validators' attestations will dominate by reputation weight in most CPMLs).
- Claims about religion.
- Claims about sexuality and gender.

The protocol does not arbitrate these. The admissibility rule is the operational criterion, not a topical-content criterion.

## Why the distinction matters

A protocol with a topical-refusal list becomes a lever of capture. Any expansion of the list — even one that seems obvious at the time — sets precedent for further expansion under political pressure. Three years of incremental expansion turns "narrow content refusal" into a de facto editorial board.

Operational refusals are harder to game. CSAM verification is not an operation anyone can plausibly argue should be permitted; the refusal is stable across political regimes. Topical refusals like "claims about vaccines" have opposite advocates on both sides and the refusal is contested from the moment it is proposed.

The foundation commits to operational-refusal-only as a bright-line architectural principle.

## The panel

### Composition

The dispute panel is a 9-seat body:

- **3 legal seats** — international human-rights lawyer, defamation specialist, content-moderation jurist. Nominated by professional associations, confirmed by foundation board.
- **3 editorial seats** — senior fact-checker (IFCN background), investigative-journalism veteran, academic ethicist. Nominated by discipline associations.
- **2 civil-society seats** — human-rights NGO representative, digital-rights advocate.
- **1 multi-stakeholder seat** — rotating from user community, elected by chapter representatives.

Seats are 3-year terms, renewable once. Staggered so no single cohort controls a panel composition. Geographic distribution required: at least 3 continents represented at all times.

### Procedures

The panel handles four kinds of matter:

1. **Expansion of the operational-refusal list.** Proposed additions require 7 of 9 supermajority + published rationale + 60-day public comment period.
2. **Contraction of the list.** Proposed removals same threshold.
3. **Appeals.** Validator, commissioner, or publisher contests a protocol-level or aggregator-level action.
4. **Systemic investigations.** On its own initiative (or by petition), the panel reviews aggregator-level patterns (e.g., a chapter's filter application).

Decisions are published in full with rationale. Dissenting opinions are published. All procedures are recorded and auditable.

### Independence

The panel is independent of the foundation board. Members are nominated and confirmed but not directed. Foundation cannot remove a panel member except for documented misconduct (per a published removal procedure with own appeal path).

Funding for the panel comes from a dedicated line item in the global treasury with its own board-ring-fenced multi-year commitment.

## The narrow initial refusal list (v0.1 — starting proposal)

1. Attestations that claim to verify the content or authenticity of child sexual abuse material.
2. Attestations that claim to verify non-consensual intimate imagery of identifiable persons.
3. Attestations that verify threats of imminent harm directed at specific identifiable persons or protected groups (as defined in ICCPR Article 20).
4. Attestations that verify detailed synthesis or operational instructions for biological, chemical, or radiological weapons usable at mass-casualty scale.
5. Attestations that claim to verify or locate active illegal markets (narcotics, trafficking, commissioned violence).

Each item is operational, not topical. Each refuses a specific operation-shape, not a topic area.

Any addition to this list requires the 7-of-9 panel supermajority described above. Any removal requires the same. The list itself is published on the protocol log as a signed foundation-board document.

## Critical analysis

**1 — The list is not short enough to be uncontroversial.** Item 4 (bio/chem/rad weapon instructions) has definitional edges. Item 3 (threats) has definitional edges. Response: these are named; the panel handles edge cases; the list is revisable.

**2 — The panel is itself a lever of capture.** A well-resourced adversary cultivates the nomination pipeline for seats; eventually the panel tilts. Response: staggered terms; geographic distribution; nomination comes from multiple independent associations, not foundation board; public dissenting opinions; independent removal procedure.

**3 — Edge cases will multiply.** Is "how to synthesise sarin using household materials" in scope for item 4, or is it fictional-drama content that happens to include harmful detail? The panel will handle, but case law takes years. Response: the protocol operates during ambiguity; refusals are reversible; the panel's precedent accumulates publicly.

**4 — Hostile-frame attestations will push the list's boundaries.** A frame chartered for "revisionist history" will produce attestations disputing facts the mainstream considers settled; it will not be refused at protocol level, only de-weighted by most CPMLs. Response: correct. The protocol's plural-verdict promise is real. Mainstream consumers see mainstream weighting; revisionist frames are recorded publicly for transparency and for the chapter / CPML composition layers to handle.

**5 — Refusal list becomes de facto editorial policy.** The foundation could be pressured to expand the list until "operational refusal" becomes a fig leaf for topical editorialising. Response: 7-of-9 supermajority + 60-day comment + published rationale + public removal right. The structural safeguards are deliberately high-friction. If the safeguards are breached, the foundation's credibility suffers immediately and visibly.

**6 — Appeals are expensive and slow.** An aggrieved validator waits months for panel review. Response: the panel has emergency procedures for urgent appeals (48-hour provisional decision by 3-member subpanel). Most appeals are not urgent; slowness is tolerable.

**7 — The panel is unfunded.** Members volunteer; work quality suffers; or they become dependent on foundation stipends that compromise independence. Response: panel members receive competitive honoraria from the ring-fenced treasury line. Amount is published; compensation is not so large as to create dependency.

**8 — Cross-jurisdictional enforcement of refusals.** A claim that the protocol refused at the log layer is still possible to post outside the protocol. The refusal binds the protocol, not the world. Response: yes. The protocol is a voluntary standard. Its refusal removes *signed attestation within this specific substrate*; actors can still publish elsewhere. This is a feature of voluntary standards, not a bug.

## Related work

- **Wikipedia's ArbCom and community sanctions.** Governance body for conduct disputes; multi-member, elected, revisable.
- **Mozilla's Community Participation Guidelines.** Narrow, specific, and enforceable standards.
- **IFCN Code of Principles.** Fact-checker professional standards; voluntary accreditation with ongoing compliance review.
- **ICANN Ombudsman.** Appeals and dispute process for internet-governance decisions.
- **National Press Complaints bodies** — per-country journalism-standards bodies; model for chapter-level dispute handling.
- **Trust and Safety Professional Association** — emerging field; practices relevant for defining what operational-refusals look like.

## Open questions

- What is the bright-line test for "universal harm"? Item 5 (illegal markets) is nearly universal; item 3 (threats against protected groups) is more jurisdictionally varied.
- Should the panel have on-chain enforcement power (pause a validator's attestation-submission ability) or is its authority purely advisory to the foundation and chapters?
- How does the panel interact with the chapter-level filters? Does a chapter's filter require panel review if it differs from the foundation's position?
- What is the appeals path for a panel decision itself? An "appeals appeals body" risks infinite regress.

## What we'd build

- **Panel charter** — governance document, written and published at Phase II launch.
- **Panel member nomination + confirmation workflow** — multi-step process with public participation.
- **`veritas-refusal-registry`** — on-chain, signed, versioned operational-refusal list.
- **`veritas-refusal-enforcer`** — client-side SDK + chain contract that rejects attestations matching refusal signatures.
- **`veritas-appeals-tracker`** — public dashboard of appeals filed, decisions, timelines.
- **Panel-member honoraria system** — via treasury; published amounts.
- **Published procedural handbook** — how the panel operates; updated annually.