RESEARCH

Effect-Transparent Governance for AI Workflow Architectures: Semantic Preservation, Expressive Minimality, and Decidability Boundaries

ArXiv cs.AI · Wed, 06 May 2026 04:00:00 GMT

arXiv:2605.01030v2 Announce Type: new Abstract: We present a machine-checked formalization of structurally governed AI workflow architectures and prove that effect-level governance can be imposed without reducing internal computational expressivity. Using Interaction Trees in Roc

Read original source Discuss with A.S.I.S