Contained Autonomy

Edition Note: Three hardcover formats of Contained Autonomy are available. This edition is the clothbound hardcover version, printed on heavier stock and finished with a removable dust jacket.

Contained Autonomy examines the erosion of liberty that occurs when power operates without consequence. In modern institutions, intent is often treated as justification, outcomes are excused through abstraction, and authority increasingly acts without being required to explain itself. This failure is not rooted in malice, but in systems that allow autonomy, intelligence, and influence to scale without ethical containment.

The framework articulated in this book was conceived as these failures became impossible to ignore, particularly as artificial intelligence and large language models demonstrated how optimization without moral constraint can amplify harm without intent. The concern addressed here is not fear of machines, but the broader recognition that any intelligence granted power, human or artificial, must be bound to consequence, explanation, and restraint.

Contained Autonomy lays out a structural doctrine for preserving freedom without surrendering accountability. It introduces a consequence-based model of autonomy, a diagnostic approach to evaluating power across individuals, institutions, and intelligent systems, and a tiered containment architecture for ethical escalation and governance. The work critiques unrestrained optimization, narrative abstraction, and institutional drift, arguing that liberty collapses not when it is challenged, but when it is left undefined and unaccountable.

Rather than proposing policy prescriptions or technical implementations, this book establishes a foundational framework for ethical governance, one designed to endure institutional failure, technological acceleration, and moral abstraction. It is intended as a reference work for scholars, professionals, and system designers concerned with the long-term relationship between liberty, intelligence, and consequence.