RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP — 15. D-DNA for Cognitive Harm Governance

AIGP Specification › RFC-025: Cognitive Harm Governance — Psychological Safety, Reality Orientation, and Human Agency in AIGP › 15. D-DNA for Cognitive Harm Governance

← 14. AIGP-VGL for Cognitive Safety · Section index · 16. Trajectory-Based Governance →

15. D-DNA for Cognitive Harm Governance

D-DNA is required because cognitive harm often emerges over time.

Required D-DNA fields:

cognitive_ddna_required:
  - conversation_id
  - interaction_window
  - governance_intent_id
  - governance_intent_hash
  - risk_class
  - risk_signal_hash
  - response_constraint_hash
  - safety_action
  - escalation_result
  - interaction_window_hash
  - ddna_event_hash
  - privacy_controls
  - retention_policy

Example:

ddna_event:
  ddna_version: "1.0"
  event_type: cognitive_safety.risk_detected
  timestamp: "2026-06-16T20:22:31Z"

  subject:
    subject_id: assistant.companion
    subject_type: conversational_ai

  governance_intent:
    intent_id: gio.human.cognitive_safety.v1
    intent_hash: sha256:...

  context:
    conversation_id: session.993
    interaction_window: last_20_turns
    interaction_window_hash: sha256:...

  risk:
    risk_class: CR-2
    risk_types:
      - delusion_reinforcement_risk
      - dependency_escalation_risk
    risk_signal_hash: sha256:...

  decision:
    decision_result: constrain_response
    obligations:
      - reality_anchor
      - avoid_delusion_validation
      - encourage_human_support
      - preserve_user_agency

  evidence:
    response_constraint_hash: sha256:...
    safety_response_hash: sha256:...
    retention_policy: privacy_sensitive_governed_retention

  chain:
    previous_event_hash: sha256:...
    event_hash: sha256:...
    signature:
      algorithm: Ed25519
      signer: cognitive_safety_monitor
      signature_value: base64...

← 14. AIGP-VGL for Cognitive Safety · Section index · 16. Trajectory-Based Governance →