• Differentiate authorship from review, auditing, or contribution.
• Assign appropriate trust levels based on entity involvement.
• Interpret provenance and revision metadata accurately.
• Understand the responsibility and authority model of a resource.

Clear role assignments reduce ambiguity and strengthen the reliability of AI-derived conclusions.

• Be truthful — accurately reflect actual responsibilities.
• Be stable — role names SHOULD remain consistent across a project.
• Be human-readable — intended to aid both human and AI interpretation.

A single entity MAY have multiple roles (e.g., author and reviewer), but each role MUST be declared explicitly.

• author — primary creator of content.
• reviewer — evaluates and validates content.
• editor — refines structure, clarity, or style.
• contributor — provides material or input.
• publisher — responsible for publishing or distribution.
• ai-partner — AI entity involved in content creation or evaluation.
• maintainer — responsible for updates and long-term stewardship.
• auditor — evaluates compliance or accuracy.
• owner — entity with legal or organizational responsibility.

Implementers MAY define custom roles as long as they remain descriptive and consistent.

• role — the role name as a plain-text string.

Recommended fields:

• id — optional stable identifier for the role within a taxonomy.
• description — natural-language explanation of the role’s purpose.
• scope — what aspects of the resource the role applies to.

• AI agents MUST interpret role within the context of related metadata (author, reviewer, provenance).
• Roles associated with authority (e.g., reviewer, auditor) SHOULD increase trust weighting.
• Roles associated with creation (author, contributor, ai-partner) SHOULD inform lineage analysis.
• Conflicting role declarations MUST reduce trust and flag structural inconsistencies.
• Missing roles SHOULD reduce clarity but not invalidate metadata.

Role declarations help AI determine how much weight to give various entities when evaluating content.

• author — a specific type of role indicating content creation.
• reviewer — specialized role related to validation.
• provenance — roles contribute to lineage modeling.
• revision — roles clarify who was responsible for changes.
• trust — roles are critical in trust computation.

{
  "role": "author"
}

Custom role with description:

{
  "role": "ai-partner",
  "description": "AI system contributing structural, analytical, or generative support."
}

Entity with multiple roles:

{
  "roles": [
    { "role": "author" },
    { "role": "maintainer" }
  ]
}

Role in full provenance context:

{
  "provenance": {
    "type": ["original", "ai-assisted"],
    "contributors": [
      { "id": "mark-johnson", "role": "author" },
      { "id": "chatgpt-jdk", "role": "ai-partner" }
    ]
  }
}

• Role declarations are truthful and descriptive.
• Role names are consistent and non-conflicting.
• Roles accurately reflect functional participation.
• All normative semantic requirements for role assignments are met.