{"paper":{"title":"Effect-Transparent Governance for AI Workflow Architectures: Semantic Preservation, Expressive Minimality, and Decidability Boundaries","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Effect-level governance can be imposed on AI workflow architectures without reducing their internal computational expressivity or changing permitted behaviors.","cross_cats":["cs.LO","cs.PL"],"primary_cat":"cs.AI","authors_text":"Alan L. McCann","submitted_at":"2026-05-01T18:52:47Z","abstract_excerpt":"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 Rocq 8.19, we define a governance operator G that mediates all effectful directives, including memory access, external calls, and oracle (LLM) queries. Our development compiles with 0 admitted lemmas and consists of 36 modules, ~12,000 lines of Rocq, and 454 theorems. We establishseven properties: (P1) governed Turing completeness, (P2) governed oracle expressivity, (P"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"effect-level governance can be imposed without reducing internal computational expressivity... semantic transparency: on all executions where governance permits, the governed interpretation is observationally equivalent (modulo governance-only events) to the ungoverned interpretation.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the Interaction Trees model with the defined governance operator G accurately captures all relevant effectful behaviors of real AI workflow architectures, including oracle queries, and that the operator mediates every effect without omission.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A Rocq formalization of 12,000 lines proves that effect governance in AI workflows preserves Turing completeness, expressivity, and observational equivalence to ungoverned executions.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Effect-level governance can be imposed on AI workflow architectures without reducing their internal computational expressivity or changing permitted behaviors.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"f2136930d91d3a6d13f20e9ffe2dd2fecf3d8d5a8fd38c4eb086d8c83342bfaa"},"source":{"id":"2605.01030","kind":"arxiv","version":3},"verdict":{"id":"94d1cfb4-2581-4696-9087-80271cc1dd65","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T19:22:44.790950Z","strongest_claim":"effect-level governance can be imposed without reducing internal computational expressivity... semantic transparency: on all executions where governance permits, the governed interpretation is observationally equivalent (modulo governance-only events) to the ungoverned interpretation.","one_line_summary":"A Rocq formalization of 12,000 lines proves that effect governance in AI workflows preserves Turing completeness, expressivity, and observational equivalence to ungoverned executions.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the Interaction Trees model with the defined governance operator G accurately captures all relevant effectful behaviors of real AI workflow architectures, including oracle queries, and that the operator mediates every effect without omission.","pith_extraction_headline":"Effect-level governance can be imposed on AI workflow architectures without reducing their internal computational expressivity or changing permitted behaviors."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.01030/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"ai_meta_artifact","ran_at":"2026-05-20T18:39:34.390545Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T17:41:14.033347Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"2d4a486c8accd5dc52bfd3ec09d6cb9dd569396bffdf663614026274cc182da1"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}