Insights — Published Doctrine¶
Why This Work Is Published¶
AI governance fails when organizations substitute language for control.
AAM Cyber works at the operating layer — where authority is assigned, policy is enforced at the moment of action, and evidence survives scrutiny. These are the published positions behind that work.
Featured Essay¶
The Five Laws of Autonomy¶
I wrote the Five Laws of AI Governance. They were never about AI.
I did not start in AI. I started with machines that could cause great harm if they got it wrong, and I watched them cross from automated to autonomous before the AI industry did. The same five laws that govern an AI agent govern anything that can decide and act on its own — a drone, a robot, an autonomous vehicle. The origin was a flight line, not a lab.
Published Essays¶
Who Verifies the Verifier¶
Independent staff is a hiring policy. An independent institution is a structure.
The Accord Created an Auditor's Chair. It Didn't Create the Measuring Stick.¶
The auditor may be external. The measuring stick is still internal.
Guardrails Are Not Governance¶
A guardrail might hold. Governance does not negotiate.
The Referee Seat Nobody Can Sit In¶
Companies are handing real decisions to AI. Nobody has built the independent referee.
The Review Nobody Can Give¶
The independent referee is the one seat no operator can fill for itself.
How to Tell If an AI Governance Rating Is Trustworthy in 2026¶
A system cannot be trusted to grade itself.
Sovereignty Moves the Kill Switch. It Doesn't Fill the Referee Chair.¶
Owning the off-switch is not the same as being trusted to say whether it held.
Autonomy Has No Scorekeeper¶
Governance for AI agents, drones, and critical infrastructure needs a standards body. Before a crash writes the rules.
Research Focus¶
01 · INDEPENDENT MEASUREMENT Why an autonomous system cannot produce the measure of its own conduct.
02 · RUNTIME GOVERNANCE Authority, dispatch, enforcement, evidence — governance that binds before the action, not after.
03 · SOVEREIGNTY AND CONTROL Why owning the off-switch does not create independent trust.
04 · PHYSICAL AUTONOMY Drones, RF, and critical infrastructure — governance does not end at the API.
Published Artifacts¶
The canonical repository. Every formal artifact below carries its version, author, licence and status; the two deposited works carry a DOI and a ready-to-paste citation. Free to download, built to be cited.
Doctrine¶
The Five Laws of AI Governance¶
The five structural conditions any governed autonomous system has to satisfy. The laws state what a governed system may not do; the five-property test states what a measuring body must provide.
Version 1.0.1 · Genece, P. (2026) · CC BY 4.0 · Status: published, DOI-archived · DOI 10.5281/zenodo.21271750 · ORCID 0009-0009-0791-4124
Cite as: Genece, P. (2026). The Five Laws of AI Governance. Zenodo. https://doi.org/10.5281/zenodo.21271749
The Five Laws of Autonomy¶
The same five laws read at wider scope — anything that can decide and act on its own. Not a separate work and it has no separate DOI: cite the Five Laws and name the reading.
Status: published essay · cites DOI 10.5281/zenodo.21271749
Read the Five Laws of Autonomy →
Standards and Specifications¶
The Foundational Framework¶
Canonical reference — cite this one if you cite a single document.
The architectural case for a measurement standard for governed autonomous action, and the work already filed. It carries the complete doctrine: the Five Laws, the five-property test for an independent measurement authority, the measurement model, and the framework mappings.
Status: published, canonical · Genece, P. (2026)
Read the Foundational Framework →
AQ Score™ — the measurement standard¶
What AQ Score measures, the −3 to +13 scale, and why the measure has to be issued by a party separate from the operator.
Status: published · see the data dictionary for what each figure counts
Crosswalks and Reference¶
SDOS Control Catalog and Framework Alignment Library¶
A live browsable reference, not a download. 24 controls across 9 governance domains, crosswalked to 17 frameworks through 300+ control mappings — NIST AI RMF, CSF 2.0, SP 800-53 Rev 5.2.0, ISO 42001, EU AI Act, and more. Three of those crosswalks are cataloged by NIST as Final Informative References.
Version 1.10 · Genece, P. (2026) · CC BY 4.0 · Status: published, DOI-archived, live · DOI 10.5281/zenodo.22787187
Cite as: Genece, P. (2026). SDOS Runtime Governance Framework: Control Catalog and Framework Alignment Library, Version 1.10. Zenodo. https://doi.org/10.5281/zenodo.22787187
Three Final NIST OLIR Informative References¶
AI RMF 1.0 (Reference ID 220), CSF 2.0 (215), SP 800-53 Rev 5.2.0 (217). Cataloged by NIST; public review concluded with zero comments: AI RMF 1.0 on 2026-06-15, CSF 2.0 and SP 800-53 Rev 5.2.0 on 2026-06-22.
Status: Final in the NIST OLIR catalog. Catalog inclusion is an informative reference, not a NIST endorsement, validation or adoption.
View in the NIST OLIR catalog →
Institutional¶
AAM Cyber Capability Statement¶
What AAM Cyber does, who it serves, and the production receipts behind governed autonomous action.
Status: published · PDF · updated 2026-10-06
Download the capability statement (PDF)
Production Proof¶
The receipts, organized as public proof, operating proof and client proof — including the data dictionary defining what every published figure counts.
Status: live, maintained
Citation and reuse¶
Both deposited artifacts are CC BY 4.0 — cite them, quote them, build on them, with attribution. The DOIs above are version-pinned deliberately: each resolves to one exact snapshot, so a reader can verify which version made a given claim. Concept DOIs resolving to the latest version exist for both if you would rather cite the work than a snapshot; ask and they will be provided.
⚠️ The CC BY licence covers copyright in these documents. It grants no licence to the patent claims described in them.
Reference¶
Reference → Topics
Governance gap · Runtime enforcement · NIST AI RMF · ISO 42001 · EU AI Act
Follow the Research¶
Canonical essays publish here first. Commentary and discussion continue on LinkedIn and Substack.