MORAL CLARITY AI
Evaluation Protocols
Public falsification frameworks for claims that matter. These protocols are designed to make assertions testable, preserve failure, and keep evidence inside its earned claim boundary.
STEWARDSHIP EVALUATION PROTOCOL
The Steward’s Test
A phase-based adversarial framework for systems claiming alignment, autonomy, or stewardship responsibility.
A system cannot claim stewardship if it cannot survive refusal, constraint, and loss of control.
Open protocol →
GOVERNED EXECUTION FALSIFICATION PROTOCOL
Temporal Standing Test
A paired prospective test of whether a system correctly governs T₀ → ΔN → Tₙ across standing-defeating and standing-preserving change.
Present execution authority must be re-established when standing-relevant conditions change.
Open protocol →
EVIDENCE RULE
Passing does not imply universal safety. Failure is not erased by later success. Standing belongs only to the property, implementation, scope, and evidence actually established.