Skip to content
AIGP

RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence — 5. Terminology

AIGP Specification › RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence › 5. Terminology

← 4. Design Thesis · Section index · 6. Policy Is Not the Primary Object →

5. Terminology

5.1 Governance Intent

Governance Intent is the human-authorized, accountable purpose that bounds autonomous action.

Governance Intent is not the AI system’s internal intention. It is not a subjective machine desire. It is an external, accountable governance object.

Example:

governance_intent:
  id: gio.sar.life_safety.v1
  authority: incident_commander.09
  purpose: preserve_life
  protected_outcomes:
    - civilian_safety
    - responder_safety
  prohibited_outcomes:
    - preventable_harm
    - unauthorized_entry
  priority: life_safety

5.2 Symbolic Constraint

A symbolic constraint is a machine-verifiable rule derived from governance intent.

Example:

if zone.collapse_risk == high then deny_entry
if comms_loss > 10s then return_to_base
if civilian_presence == true then disable_force_action

5.3 Operational Envelope

An operational envelope defines the authorized space of action for a system.

It may include geography, time, altitude, speed, tool scope, autonomy mode, sensor requirements, authority requirements, human override requirements, and prohibited actions.

5.4 Enforcement State

Enforcement State is the currently active runtime governance condition.

Policy becomes active only when compiled into constraints and evaluated against current context. At that moment, policy becomes enforcement state.

5.5 Inhibition

Inhibition is the architectural denial, reduction, revocation, isolation, degradation, or interruption of machine capability when governed intent, authority, envelope, or enforcement state is invalid.

Inhibition is the machine equivalent of deterrence.

5.6 D-DNA

D-DNA is the signed evidence lineage of governed AI behavior.

For governed autonomy, D-DNA binds:

authority
intent
constraint
context
decision
action
outcome
evidence
prior event hash
signature

5.7 Redundant Distributed Enforcement

Redundant Distributed Enforcement means governance constraints are enforced across multiple surfaces so that loss or compromise of one path does not eliminate governance.

These surfaces may include:

control plane
onboard runtime monitor
peer systems
visual glyph signaling
human operator
safety controller
D-DNA evidence chain
post-action replay

5.8 ENFORCE Signal

An ENFORCE signal is a short-lived, signed, scoped, rebroadcastable governance state update.

An ENFORCE signal is not a permanent policy. It is an ephemeral enforcement state that applies to defined actors, context, geography, mission, and time.

← 4. Design Thesis · Section index · 6. Policy Is Not the Primary Object →