Tracing AI Action Back to Human Authority: Authority Provenance, Authorization, and Revocation in Microsoft AI's Humanist AI Code of Conduct

Tracing AI Action Back to Human Authority applies Institutional Authority Dynamics (IAGD) to Microsoft AI’s draft Humanist AI Code of Conduct. The analysis accepts Microsoft’s existing Chain of Command and Human Control architecture as a substantive governance foundation and examines a narrower problem: whether consequential AI action can remain traceable to the human authority from which its authorization derives as execution moves through models, agents, tools, systems, and downstream delegations. The intervention introduces Authority Provenance as a complementary governance requirement for reconstructing the lineage of authorization from human authority through action, scope, conditions, delegation, transformation, execution, correction, revocation, and current standing. It distinguishes a Chain of Command—which determines whose instructions govern—from Authority Provenance, which determines whether the authority supporting a consequential action can be reconstructed from source to execution and back through correction, revocation, or renewal. The analysis further examines multi-agent governance, downstream authorization, revocation propagation, Authority-Control Alignment, and the distinction between action logging and authority lineage. It recommends that consequential AI actions remain attributable through a human-legible authorization lineage and proposes an Authority Provenance Reconstruction evaluation for testing origin, scope, lineage, authorization standing, correction rights, and revocation propagation. This record is an Institutional Authority Dynamics Policy and Technical Application Record. It is non-canonical and represents a bounded application of IAGD. Application does not constitute empirical validation, Microsoft adoption, Microsoft endorsement, or activation of the BlackGuard Multilevel Governance Architecture.