Why AI Governance Frameworks Cannot Deploy

The AI governance conversation has produced dozens of frameworks, hundreds ofpublished papers, and an entire conference circuit. Almost none of it has deployed.This paper argues that non-deployment is not a sign of technical immaturity. It isthe predictable structural outcome of designing governance systems that require theentities with the most to lose from transparency to voluntarily adopt transparency.The frameworks describe what should happen at the execution boundary. They donot account for the fact that the execution boundary is owned by the entity thatbenefits most from keeping it opaque. Until governance enters through a compliancevector that cannot be routed around — rather than a voluntary adoption pathwaythat can — the frameworks will continue to not deploy. Post-quantum cryptographictransition mandates represent the most credible current forcing function.