Founder & CEO, FERZ, Inc. Building runtime authorization infrastructure, the enforcement layer of AI governance, for regulated industries.

ORCID: 0009-0008-8012-6100

Published, vendor-neutral work on AI governance that can be independently verified, not merely documented and trusted.

Core thesis: AI governance is an enforcement problem, not a monitoring problem. Governance must shift from policy statements and process documentation to authorization artifacts that support independent replay. If you can't replay the decision, you can't audit it. If you can't audit it, you can't govern it.

This approach treats governance as a runtime property of the system, not an after-the-fact reporting function.

The work centers on tamper-evident authorization artifacts, implemented as Proof-Carrying Decisions under 5TS: artifacts that carry verification evidence, enabling fail-closed authorization where effect-bearing actions do not execute without a verdict.

The Five Tests Standard (5TS) is a published, vendor-neutral standard defining five conformance tests for AI governance mechanisms. It supersedes the Four Tests Standard (4TS); the repository name is retained from the predecessor.

• Stop. The system can be halted before side effects occur.
• Ownership. Each consequential decision maps to a named accountable authority.
• Replay. The decision can be reproduced at the boundary.
• Escalation. Control transfers at defined policy boundaries.
• Provenance. The inputs grounding a verdict have an established origin. Origin, not truth.

The verdict space is ALLOW, DENY, and ABSTAIN. An ABSTAIN blocks execution pending authorized human override; escalation is the consequence of ABSTAIN, not a fourth verdict.

Archived under concept DOI 10.5281/zenodo.21040295.

A decision-grade guide series for executives, regulators, and system owners responsible for deploying AI in regulated environments.

The series translates deterministic governance principles into actionable authorization models, clarifying:

• What must be verifiable vs. what can be merely documented
• Where traditional compliance frameworks stop short of pre-execution authorization evidence
• Why logs, audits, and attestations are insufficient as authorization evidence
• How to operationalize fail-closed authorization for AI systems

Written for leaders who must authorize AI systems, not merely observe them.

• 📚 SSRN: https://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=7471418
• 🧾 Zenodo: https://zenodo.org/communities/ferz/
• 🔬 ResearchGate: https://www.researchgate.net/profile/Edward-Meyman
• 🛡️ IP.com: defensive publications establishing AI governance prior art; search: Meyman

These publications establish architectural and conceptual prior art for deterministic AI governance, authorization artifacts, and independently reconstructable verification.

Patent-pending governance infrastructure addressing distinct failure modes in AI authorization, escalation, and accountability under regulatory constraints.

Patent filings and published prior-art materials include work on:

• LASO(f): deterministic linguistic policy enforcement
• DELIA: deterministic governance of AI-generated multimodal outputs and serialized action descriptions
• CausaCore: causal governance and decision lineage infrastructure

The filed portfolio is complemented by defensive publications establishing foundational prior art. Mechanism details are not published outside the filings themselves.

• The standard defines what must be verifiable
• The systems enforce authorization at runtime
• The publications explain why observation-based approaches fail

Governance architecture, not product documentation.

• Standards adoption: organizations implementing verifiable governance frameworks
• Pilot customers: regulated enterprises ready to move beyond compliance theater
• Collaborators: researchers and engineers working on deterministic governance, formal methods, or regulatory technology

• 🌐 Website: https://ferz.ai
• 📧 Research correspondence: edward@ferz.ai
• 📧 FERZ and pilot inquiries: info@ferz.ai
• 💼 LinkedIn: https://www.linkedin.com/in/edmeyman/

Governance without verification is theater. Verification without determinism is sampling. — Edward Meyman