This white paper extracts a minimum viable audit profile (MVAP) from the broader Hierarchical Physical Compliance Architecture (HPCA) family and adapts it for EU AI Act Article 14 human oversight. The core proposition is that high-risk AI oversight cannot rely solely on software logs, model cards, or post-hoc semantic records when the contested event may itself have been synthetically generated, replayed, or physically bypassed. HPCA introduces a physical-causality layer that binds oversight, interruption, and forensic evidence to measurable thermodynamic, spatial-inertial, and hardware end-state observables. Version 1.6 strengthens the AI usage disclosure language, explicitly names the Thermodynamic Energy-Tariff, the Irreversible Physical Severance Architecture, the Non-Markovian Biological Phase Operator, and the Hierarchical Physical Compliance Architecture as exclusive original works of the human author, and restores the funding declaration and Prior Art Non-Admission Clause.
Chin-Yu Hsu (Thu,) studied this question.