What question did this study set out to answer?

February 20, 2026Open Access

Execution Admissibility Theory for Autonomous Agents — Part II: Control Surface and Regime Separation

AKAslanbek Yurievich KhamikoevChinese Academy of Governance

Key Points

This part aims to formalize a control-surface framework and regime separation model for autonomous agents.
Introduced control-surface formalization
Established policy-cone relationship
Developed execution regime separation
Defined control-surface integrity criteria
Outlined insurability conditions as regime properties
Established formal specifications for audit and regulatory contexts
Clarified the relationship between control surfaces and regime properties
Provided theoretical insights into non-implementation scopes

Abstract

This record contains Part II of the Execution Admissibility Theory for Autonomous Agents series. It is a formal theoretical specification intended for audit, regulatory, and insurability evaluation contexts. Title: Execution Admissibility Theory for Autonomous Agents — Part II: Control Surface and Regime Separation Author: Aslanbek Yurievich KhamikoevIndependent Researcher, AI Systems Governance and Execution Safety Position in Series Part II develops the structural control-surface framework and regime separation model introduced after the bounded-loss execution results established in Part I. Part I (bounded-loss theorem anchor) establishes formal deployability and insurability conditions under invariant-mediated execution-admissibility constraints. Part II introduces: Control-surface formalization Policy-cone relationship Execution regime separation Control-surface integrity criteria Insurability condition as regime property Audit and regulator interpretation layer Explicit non-implementation scope boundary This document is theoretical and property-level. It does not define protocol formats, cryptographic constructions, or machine state machines. Scope Classification Document role: Theory / Regime Model / Audit and Regulator Interpretation Layer Status: Non-Implementation Theoretical Specification Applicability: Effectful execution channels only Relation to Series Supplement to: Execution Admissibility Theory for Autonomous Agents — Part I: Deployable and Insurable AGI Requires Verifiable Execution Admissibility (DOI: 10. 5281/zenodo. 18663653) Foundational basis for planned Part III: Execution Admissibility Theory for Autonomous Agents — Part III: Non-Bypass and Conformance Theory Files Included in This Record ExecutionAdmissibilityTheoryPartII. pdfCanonical formatted publication version ExecutionAdmissibilityTheoryPartII. texLaTeX source (Overleaf-ready) ExecutionAdmissibilityTheoryPartII. txtPlain-text archival version File order reflects canonical publication hierarchy: (1) formatted publication record, (2) source document, (3) archival plain-text reference. This series forms a structured DOI-linked theoretical corpus on execution admissibility as a deployability and insurability condition for high-capability autonomous systems.

AIに質問

Bookmark

View Full Paper