Beyond Oversight: A Deterministic Dual-Boundary Control Architecture for Automated AI R&D Under Intelligence-Explosion Conditions

This paper presents a deterministic dual-boundary control architecture for automated AI research and development under conditions in which AI-enabled R&D cycles may become substantially faster than ordinary human and institutional response cycles. The paper is a structural response to the 2026 Cambridge Programme on AI Science & Policy report, “What if automating AI R&D triggers an intelligence explosion?” It focuses on the control-latency problem that arises when visibility, reporting, auditing, and human oversight remain necessary but may no longer be fast enough to constitute the primary runtime control surface. The proposed architecture separates two control domains. The first is the Human Authority Boundary, which preserves human discretion, responsibility, acceptance, and institutional authority without treating AI-generated outputs as self-authorizing decisions. The second is the Machine-Enforceable Control Boundary, which constrains protected state transitions through explicit authorization conditions, independent verification, evidence-lineage preservation, scoped commit boundaries, revocation, failure locality, recovery controls, and fresh readmission requirements. The paper develops machine-verifiable control invariants, a reference control path for automated AI R&D, a security-of-security model for AI systems that monitor other AI systems, adversarial failure modes, evaluation protocols, deployment scenarios, and standardization implications. The central distinction is that observation is not authorization, evidence is not authority, verification is not execution, human acceptance is not equivalent to technical determinacy, and recovery is not equivalent to restoration of prior authority. This publication is intentionally limited to public research-level abstractions. It does not disclose unpublished patent claim language, confidential claim charts, private source locators, provider-specific production parameters, or non-public implementation details. Structural Paper Series - Paper 01 Final Publication Edition v2.1 Research Program on Deterministic Infrastructure and Human-Centered AI Coordination Transition Intelligence Institute, Switzerland Institutional Establishment in Preparation Foundational Research Signature: The Second Waters