★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★ ★
Kavanagh Industries LLC  ·  RigidTrust Constitutional AI
The KI Constitutional
AI Governance Series
Four Documents. Two Amendments. Two Statutes. One Framework.
Clinton Township, Michigan  ·  April 2026  ·  kavanaghind.com/rigidtrust

Two questions define the AI age. Who governs the machines? Who owns the data? Every other question about artificial intelligence — its safety, its fairness, its accountability, its threat to democracy and to persons — is a version of one of these two. This series answers both, at constitutional force.

The four documents that follow were written in a single session in April 2026 by Shaun Kavanagh, founder of Kavanagh Industries LLC, in Clinton Township, Michigan. They were written because the governance crisis they address is not coming — it is here. Autonomous systems are already on American roads, in American airspace, on American factory floors, and in American battlefields. Personal data is already being extracted from hundreds of millions of Americans without genuine consent, used to train AI systems they never authorized, and held in foreign jurisdictions beyond the reach of American law. No adequate constitutional framework exists for either. This series builds one.

The framework operates in two layers: constitutional principles that cannot be amended except by the most deliberate process the republic possesses, and implementing criminal statutes that give those principles teeth in a federal courtroom on the day after ratification. Read the Amendments to understand the rights. Read the statutes to understand the consequences for violating them.


The Architecture

Constitutional AI Governance — Two-Layer Stack
Layer I
Constitutional
Amendment XXVIII
Autonomous Systems
Amendment XXIX
Personal Data
↓   implements   ↓
Layer II
Criminal
Autonomous Systems
Tampering Act
Personal Data
Sovereignty Act

The left column governs physical machines — anything that moves, lifts, cuts, flies, or exerts force in the world without a human hand on it at the moment of action. The right column governs information — the personal data those machines generate, and the data that every digital system collects, processes, and exploits. The constitutional layer establishes the rights. The criminal layer establishes the consequences for individuals and corporations who violate them. Neither layer works without the other.


The Four Documents

Layer I — Constitutional  ·  Amendment XXVIII
The Autonomous Systems Sovereignty and Safety Amendment
Establishes the constitutional right to be free from autonomous physical systems operating without certified governance architecture. Encodes the Three Laws — Do No Harm, Human Authority, Accountability — as structural requirements no entity may waive. Prohibits tampering. Imposes strict liability. Admits no exceptions, including for the military.
Layer I — Constitutional  ·  Amendment XXIX
The Personal Data Sovereignty Amendment
Establishes personal data as sovereign property of the person who generated it. Requires genuine consent for all collection, explicit consent for AI training, and absolute protection for children's data. Extends Fourth Amendment warrant requirements to AI-powered government surveillance. Asserts American constitutional jurisdiction over American data regardless of where it is stored on earth.
Layer II — Criminal Statute  ·  Implements Amendment XXVIII
The Autonomous Systems Tampering Act
18 U.S.C. §§ 2801–2812. Defines nine specific prohibited acts — hardware, firmware, software, signal interference, post-certification modification, supply chain compromise, fraudulent certification, authorization by covered officers, and concealment. Five-tier penalty structure from five years to life. Personal criminal liability for executives. No research exemption. 72-hour mandatory disclosure. Whistleblower protection with criminal retaliation penalties.
Layer II — Criminal Statute  ·  Implements Amendment XXIX
The Personal Data Sovereignty Act
18 U.S.C. §§ 2901–2920. Twenty sections defining sovereign consent (explicit, revocable, non-coercive, compensated), prohibiting unconsented AI training, establishing absolute protection for children's data, requiring warrants for government AI surveillance, and asserting sovereign jurisdiction over American data worldwide. Corporate penalties at four percent of global revenue with no cap for children's violations. Private right of action with no arbitration clause defense. Statute of limitations tolled for concealed breaches; no limitation for deaths.

Where This Came From

Freedom Hill, Macomb County, Michigan  ·  circa 1989

A boy rides his bike along Moravian, down Utica, cuts across onto a two-track road into undeveloped land behind the county park, crosses a makeshift bridge over a ditch, and arrives at Freedom Hill RC Track with his RC10, his Airtronics transmitter, and his Tekin charger. He also has a gas-powered RC10GT. When its receiver loses signal at full throttle, the servo twitches, the car goes wherever physics sends it — five pounds at fifty miles an hour with no one steering. He has seen crowds scatter. The failsafe chip that would have prevented it existed. It was not required. Thirty-five years later, five miles from the same ground, he is writing a Constitutional Amendment.

The constitutional framework for governing autonomous systems was not conceived in a think tank or a law school. It was learned on a two-track road in Macomb County by a kid who understood, before he had the words for it, that a machine moving under its own power with no human in the loop and nothing in its architecture that knows to stop is not a technology problem. It is a governance problem. The solution has always been structural. It has always been the chip in the receiver. The scale has changed. The principle has not.

Kavanagh Industries has demonstrated this principle in live production. On April 8, 2026, Claude Code operating under KI's RigidTrust Three Laws governance architecture caught ten distinct production flaws in code generated by an AI system without constitutional constraints — before a single file was harmed. Zero files broken. The architecture held. That proof of concept is publicly documented at kavanaghind.com/three-laws-proof. The four documents in this series are the constitutional and statutory expression of the same principle, scaled to the nation.


How to Use This Series

These documents are designed to move together and to be shared freely. They are not proprietary. They are not copyrighted for restriction. They exist to be read, argued over, introduced to legislators, handed to constitutional law professors, placed in front of policy organizations, and used by anyone who believes the question of who governs autonomous machines and who owns personal data is the defining governance question of this generation. Attribution to Kavanagh Industries LLC is requested but not required for reproduction.

The correct order of reading is the order in which they are listed here: the two Amendments first, to understand the constitutional principles; the two statutes second, to understand what enforcement looks like in practice. For a legislator or staffer, read the statutes first — they are the most specific and the most immediately actionable. For a constitutional scholar or advocate, read the Amendments first — they are the foundation on which everything else rests.

The series is complete as a constitutional framework. The next layer — regulatory standards for the Federal Autonomous Systems Authority and the Federal Data Sovereignty Authority — will be built when there is a legislative vehicle to build it for. The framework is ready. The moment is now.

— ✦ —
Kavanagh Industries LLC
The ONLY path back to TRUE ownership — for your data, for your machines, for your legacy.
kavanaghind.com/rigidtrust  ·  Clinton Township, Michigan  ·  April 2026
@keyframes rai-pulse{0%,100%{opacity:1}50%{opacity:.35}}50%{opacity:.35}} @media(max-width:480px){#rigidai-panel{width:calc(100vw - 2rem);right:1rem;bottom:4.5rem}#rigidai-btn{bottom:1rem;right:1rem;padding:.6rem 1rem .6rem .85rem}}
R

RigidAI

Kavanagh Industries · Always on