Abstract
This publication discloses a governance signal taxonomy defining six semantic classes of observable events meaningful for governance purposes — Detection, Awareness, Action, Verification, Authority, and Outcome — and a corresponding typed lifecycle object model for managing governance signals, obligations, controls, risks, and outcomes as first-class governance artifacts with defined lifecycle states and downstream consumption interfaces. The disclosed taxonomy classifies events by the governance function they serve, which is orthogonal to security severity classification used in conventional SIEM systems and to timestamp-actor recording used in conventional audit trails. In one embodiment, a governance operational layer receives runtime events from heterogeneous sources, evaluates each event against a formal governance ontology defining admissible signal classes and required structural fields, admits conformant events as typed governance objects, produces structured absence records for required signal types not present within a closed observation window, and maintains a rejection log with chained proofs such that insertion, deletion, or reordering of received events is detectable. The layer is designed to be independently operable and produces governance-typed runtime objects regardless of which downstream evaluation, notarization, or execution authorization systems consume them.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Pellicano, Roy, "Governance Signal Taxonomy, Typed Lifecycle Object Model, and Semantic Classification Framework for Runtime AI Governance Infrastructure", Technical Disclosure Commons, (April 16, 2026)
https://www.tdcommons.org/dpubs_series/9804