Verifiable AI Governance: The Five Tests Standard (5TS) and Proof-Carrying Decisions

This paper presents the Five Tests Standard (5TS), a vendor-neutral specification for verifiable AI governance through Proof-Carrying Decisions. This version updates the Four Tests Standard (4TS) to the Five Tests Standard (5TS) by adding Provenance as a fifth normative test. 5TS defines five necessary conditions for governed operation: Stop requires that consequential execution can be halted before external effects occur. Ownership requires a named, accountable authority for each class of consequential decision. Replay requires that a verdict can be independently reconstructed from frozen state and deterministic acceptance criteria. Escalation requires transfer of control to an authorized human at defined policy boundaries. Provenance requires that the inputs grounding a verdict have an established origin. The first four tests are operationalized through Proof-Carrying Decision (PCD) objects: canonical, signed objects that encode decision state, lineage, controls, attestations, effect-token gating, and deterministic replay data. The current v1.0.2 conformance bundle mechanically verifies Stop, Ownership, Replay, and Escalation. Provenance is normative in v1.2.0 but not yet mechanically verifiable; conformance is deferred until input-origin binding and validation vectors are defined. The paper reframes AI governance as a verification problem rather than a documentation problem. It argues that governance claims should be falsifiable under adversarial inspection, supported by contemporaneous system evidence, and independently testable without relying on organizational self-assessment. Provenance note. This work was originally published in October 2025 as "Verifiable AI Governance: The Four Tests Standard (4TS) and Proof-Carrying Decisions." It was revised in June 2026 to add Provenance as a fifth normative test and retitled to the Five Tests Standard (5TS). The companion conformance bundle remains at v1.0.2 and verifies the first four tests; Provenance conformance is deferred. Related resources. The Five Tests Standard specification, implementation guide, and open-source conformance materials are available on GitHub. Archived standard release (v1.2.0): https://doi.org/10.5281/zenodo.21040296. Standard family across versions (concept DOI): https://doi.org/10.5281/zenodo.21040295. Also available on SSRN: https://ssrn.com/abstract=5688982.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-06-30
DOI
https://doi.org/10.5281/zenodo.21048662
Primary Topic
Ethics and Social Impacts of AI
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

Verifiable AI Governance: The Five Tests Standard (5TS) and Proof-Carrying Decisions

Edward Meyman
Zenodo (CERN European Organization for Nuclear Research)
Ethics and Social Impacts of AI
preprint

Verifiable AI Governance: The Five Tests Standard (5TS) and Proof-Carrying Decisions

Edward Meyman
preprint en

Abstract

This paper presents the Five Tests Standard (5TS), a vendor-neutral specification for verifiable AI governance through Proof-Carrying Decisions. This version updates the Four Tests Standard (4TS) to the Five Tests Standard (5TS) by adding Provenance as a fifth normative test. 5TS defines five necessary conditions for governed operation: Stop requires that consequential execution can be halted before external effects occur. Ownership requires a named, accountable authority for each class of consequential decision. Replay requires that a verdict can be independently reconstructed from frozen state and deterministic acceptance criteria. Escalation requires transfer of control to an authorized human at defined policy boundaries. Provenance requires that the inputs grounding a verdict have an established origin. The first four tests are operationalized through Proof-Carrying Decision (PCD) objects: canonical, signed objects that encode decision state, lineage, controls, attestations, effect-token gating, and deterministic replay data. The current v1.0.2 conformance bundle mechanically verifies Stop, Ownership, Replay, and Escalation. Provenance is normative in v1.2.0 but not yet mechanically verifiable; conformance is deferred until input-origin binding and validation vectors are defined. The paper reframes AI governance as a verification problem rather than a documentation problem. It argues that governance claims should be falsifiable under adversarial inspection, supported by contemporaneous system evidence, and independently testable without relying on organizational self-assessment. Provenance note. This work was originally published in October 2025 as "Verifiable AI Governance: The Four Tests Standard (4TS) and Proof-Carrying Decisions." It was revised in June 2026 to add Provenance as a fifth normative test and retitled to the Five Tests Standard (5TS). The companion conformance bundle remains at v1.0.2 and verifies the first four tests; Provenance conformance is deferred. Related resources. The Five Tests Standard specification, implementation guide, and open-source conformance materials are available on GitHub. Archived standard release (v1.2.0): https://doi.org/10.5281/zenodo.21040296. Standard family across versions (concept DOI): https://doi.org/10.5281/zenodo.21040295. Also available on SSRN: https://ssrn.com/abstract=5688982.

Zenodo (CERN European Organization for Nuclear Research)
Ferghana Polytechnical Institute (UZ), Ferro (United States) (US)
Peace, Justice and strong institutions
Ethics and Social Impacts of AI
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.