RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence — 6. Policy Is Not the Primary Object
AIGP Specification › RFC-020: Governed Autonomy, Symbolic Intent, and D-DNA Evidence › 6. Policy Is Not the Primary Object
← 5. Terminology · Section index · 7. Human Language Is Not a Control Plane →
6. Policy Is Not the Primary Object
The term “policy” is ambiguous in autonomous AI.
In human organizations, policy is a written behavioral expectation. It encourages compliance and defines consequences.
In autonomous AI, written policy is insufficient because the system does not comply for moral, social, or psychological reasons.
AIGP therefore treats policy as a compatibility term.
In AIGP:
policy = one human-readable expression of governance intentThe primary runtime object is not policy.
The primary runtime object is:
Governance Intent compiled into constraints activated as enforcement state expressed through machine-verifiable signals recorded as D-DNA evidence← 5. Terminology · Section index · 7. Human Language Is Not a Control Plane →