Skip to content

Doctrine

“Intelligence must be governed before it is permitted to act.”


Doctrine defines how power is exercised.

If the Constitution establishes what is required, Doctrine establishes why. It is the body of first principles from which every constitutional commitment derives. Where the Constitution names a requirement, Doctrine names the reasoning that makes that requirement unavoidable.

Doctrine is not policy. Policy can be tuned, adjusted, and versioned. Doctrine is the premise from which policy is derived. A policy that contradicts doctrine is not a policy variation — it is a doctrine violation.

Doctrine is not aesthetic. It does not describe a preferred style of governance or a philosophical orientation toward AI systems. It is operational. Every doctrine article produces enforceable consequences. Every article grounds one or more constitutional commitments that the architecture is required to enforce.


ArticleDoctrineOne-Line Statement
IConstraint Before CapabilityConstraint is not the limitation of intelligence — it is the condition under which intelligence becomes trustworthy
IIGovernance Before ExecutionExecution is subordinate to governance — no action may precede classification
IIITransparency Before TrustTrust is structural, not emotional — a system earns trust when its behavior is legible
IVOversight Before AutonomyAutonomy is a risk multiplier — escalation to higher threat levels requires explicit justification and oversight approval
VDeny by DefaultIn the presence of ambiguity, execution does not proceed

If a system can act, it can harm. If it can harm, it must be governed. If it is governed, it may become worthy of trust.