Why This Book Exists

For roughly three centuries in the Western legal‑bureaucratic tradition, sovereignty rested on a single architectural assumption: the State occupied the highest vantage point. From that altitude, it could observe territory, classify activity, and issue commands that flowed downward through the transparent medium of law. Visibility and authority were inseparable; to see was to govern.

That architecture no longer holds.

We have crossed the threshold from the Information Age — when computers processed data for humans to decide — into the AI era, where software acts on its own. AI systems now classify, recommend, negotiate, allocate, and enforce at scales and speeds that exceed the observational capacity of any sovereign. The State has not lost its legal authority; it has lost its physics. Its inherited tools now operate under physical and informational limits they were never designed to meet.

This book is a systematic diagnosis of that condition — and a structural proposal for what may succeed it.

It is also a versioned release. Because AI capabilities and regulatory frameworks evolve faster than traditional publishing cycles, this manuscript is issued as Code After: v0.9 (Open Access Edition) — an interim release intended to enter scholarly and policy debate while the full edition is prepared.

One scope clarification is necessary. This analysis focuses on the oversight of centralised, cross‑border AI systems — the computational platforms operated by frontier AI labs, hyperscalers, major technology firms, and multinational organisations through which regulatory obligations are translated and enforced. Open‑weight models introduce a parallel dynamic, extending AI activity beyond centralised compliance perimeters. Their regulatory treatment will be addressed in the full edition.

The Argument

The argument unfolds in five movements.

Part I establishes the structural problem. It introduces the Visibility Gap and the RuleExecution Gap, showing why inherited legal and oversight frameworks cannot regulate AI systems whose operations are opaque, high-velocity, and increasingly agentic.

Part II examines law and accounting as the dual constitutive languages through which economic reality is defined, measured, and governed. AI systems — like corporations and financial instruments before them — become legible to systems of control only when translated into these institutional grammars. The analysis identifies the Categorical Gap and the Measurement Gap, develops the Invisible Charter as the fused evaluative grammar through which AI systems become economically real, and shows how interpretive authority has migrated to professional intermediaries who now function as the Translation Layer.

Part III traces the migration of governance from horizontal, border-based frameworks to vertical, stack-based architectures. It introduces the AI Governance Stack, the Agentic Shift, and the emergence of vertically integrated AI ecosystems, situating the United States (U.S.), China, and the European Union (EU) as three increasingly distinct institutional regimes whose divergence makes cross-border translation structurally necessary.

Part IV introduces the Translation Layer as operational machinery — beginning with Hong Kong as a reference implementation of the institutional infrastructure through which incompatible governance stacks currently communicate. From that structural analysis, it develops the Incorporation Heuristic as the logic shaping which rules travel, stabilise, or fail within globally networked markets, and presents the Sovereignty Paradox, the Gateway
Rules framework, and the AI Power Hierarchy as structural consequences of translation under conditions of opacity and national competition.

Part V begins with the physical foundations of AI — the material constraints imposed by compute concentration, energy demand, and non-territorial strategic assets that shape national agency. It diagnoses the Legitimacy Crisis: the democratic deficit in the governance arrangements now operating. It concludes with the Third Architecture — the Sovereign API, Parametric Democracy, and the Protocol of Federation — as institutional designs for governing what follows.

The manuscript advances three contributions. First, it reframes AI governance by shifting the analytical centre from model behaviour to the institutional systems that render AI visible, legible, and governable. Rather than treating AI as a technical object to be regulated, the analysis shows how law, accounting, and administrative practice jointly construct the economic reality in which AI operates. Second, it advances accounting theory by demonstrating that the core operations of recognition, measurement, classification, disclosure, and consolidation no longer align when applied to probabilistic, networked, and emergent systems, producing a structural Measurement Gap that limits states' capacity to perceive, value, and govern AI-driven economic activity. Third, it develops a constitutional analysis of how governance functions when intelligence becomes distributed. The analysis formalises which rules propagate across borders and why, maps the resulting hierarchy of operational influence, and proposes institutional mechanisms through which democratic accountability can be extended to systems that currently operate beyond its reach.

The Stakes

AI is not unmanaged. It is shaped by default institutions, through mechanisms no one designed directly and without meaningful democratic consent. Professional services firms —
the global law firms and the Big Four accounting networks — have quietly become the operational stewards of AI governance, not because they sought this role but because the
structure of the AI era placed it on their desks. They convert legal requirements into operational systems and produce the templates that function as AI’s invisible legislation.
This arrangement sustains a functioning global AI economy across a fractured world, but it carries a democratic deficit. No one voted for the Big Four, and no constitution imagines
global law firms writing the operational rules for machine intelligence. This book makes that machinery visible, clarifies the available design choices, and proposes the institutional
architecture that may succeed it.

Why a New Constitutional Framework Is Needed

AI governance does not fail for lack of principles. Ethical frameworks proliferate, national strategies multiply, and regulatory proposals accumulate. The constraint is more fundamental: rules do not execute themselves.

Between a legislative instrument and AI behaviour lies a structural gap. Law is abstract, territorial, and deliberative; computation is specific, distributed, and instantaneous. Effective authority depends on the machinery beneath it — the operational procedures, compliance architectures, and technical controls through which rules become executable inside AI ecosystems. This requires design, not aspiration. It requires translators capable of conversion and a constitutional framework built for the operating realities of the AI era rather than the assumptions of earlier regimes.

A Note on Authorship

This book is written by a single author, but it did not emerge from a solitary path. It began as a shared inquiry with my daughters, Natarsha and Jacqueline, as they entered their studies in law and accounting at Berkeley Law and NYU Stern. Their questions, research contributions, and critical engagement — spanning constitutional structure, institutional design, accounting standards, and the mechanics of global practice — shaped the conceptual terrain of this project and helped reveal where doctrine no longer maps cleanly onto operational reality.

This book reflects that shared journey. It is not a co-authored work, but it is collaborative in origin, grounded in the belief that scholarship is a long arc built across generations. The conceptual frameworks developed here are intended as a foundation for independent scholarship — and my hope is that Natarsha and Jacqueline will extend, challenge, and build upon them in their own right.