HALE: Hybrid AI/LLM Enforcement — A Split-Responsibility Architecture for Deterministic Governance of Autonomous AI Agents

A governance architecture that pairs a predictive LLM for semantic tier classification with deterministic code enforcement for all binary decisions — combining the intelligence of a judge with the consistency of law