An academic benchmark on long-horizon agentic AI failure (HORIZON, arXiv 2604.11978) never mentions SOURCE 0 — its doctrine was published two months later. Yet its own engineering recommendation and a documented real-world incident converge, independently,

Background-agent architectures promise to fix low AI adoption by moving humans to occasional review. That review becomes the sole surviving record of human diligence — and nothing currently proves it happened.

Anthropic's own review of a Claude access incident is credible — and, on its own, unfalsifiable by anyone outside Anthropic. This is the Endogenous Audit Paradox in its most literal form.

Article 14 of the EU AI Act requires high-risk systems to be stoppable by human oversight — a capability obligation, never a timing proof. The 21 July 2026 OpenAI/Hugging Face containment failure, reported by Reuters, shows why that distinction has direct financial exposure under Article 99.