Prove human.
Reveal nothing.
The objective is not to make everyone identify themselves everywhere. The objective is to prove the minimum fact a service needs: a legitimate human root exists behind an action or an authorized agent.
A human can authorize an agent
without surrendering control.
Human proof is only the root of the decision. Permission Zero then separates the autonomous actor from the person behind it and records the exact authority that was delegated. The agent can act independently only inside that boundary.
Human root. Establish the accepted authority root using the minimum proof required by the service.
Agent identity. Bind the autonomous actor to a distinct identifier and the root that delegated authority to it.
Scoped permission. Define the permitted action, resource, value, time window and any additional policy conditions.
Revocation. End authority immediately when the owner or policy requires it; past identity does not create permanent access.
Superintelligence should not mean super-authority.
An agent may reason beyond its owner in a task and still remain bounded by that owner’s authority. Intelligence and authority are separate properties.