Execution-Boundary Interlocks for High-Autonomy AI Systems

Abstract

As AI systems transition from advisory tools to autonomous actors with execution authority, the primary risk surface shifts from model misalignment to authority misconfiguration. Existing safety discourse largely focuses on model behavior, alignment techniques, and output filtering. However, real-world harm increasingly arises from over-permissioned integrations, insufficient revocation mechanisms, and weak execution boundaries. This paper proposes Execution-Boundary Interlocks (EBI) as a governance primitive for high-autonomy AI systems. EBI is a non-bypassable architectural mechanism at the action layer that enforces scoped authority, tiered autonomy limits, deterministic revocation, and reconstructable accountability independent of model output. The framework introduces measurable governance capacity, tiered autonomy control, regulatory threshold alignment, authority segmentation, and drift detection as structural invariants for sustainable AI deployment. Rather than constraining intelligence, EBI constrains authority. This distinction enables scalable, interruptible, and accountable AI systems across cloud, enterprise, open-source, and multi-agent deployments.

Keywords