Society & Economicsarticle2026-08-23

The Structural Law of AI Systems Under Transformation - Artificial Intelligence: A Canonical Reference for Persistence, Development, Agency, Alignment, and Governance

Open access0 citations

Abstract

The Structural Law of AI Systems Under Transformation applies LA PROFILÉE (LP) to artificial intelligence. LA PROFILÉE is proposed as a structural law of systems under transformation, derived from the structure and behavior of natural systems. This work examines what follows when that structural law is applied to AI systems. The application extends across the structural life of artificial intelligence: system formation, persistence, development, structural limits, identity, transformation, transmutation, succession, agency, memory, self-modification, multi-agent organization, alignment, safety, and governance. It asks how an AI system becomes a determinate system, how it can change while remaining itself, where its structural limits lie, and what happens when those limits are crossed. Governance is therefore not the foundation of the work but one consequence of the structural architecture. The same analysis determines what system is acting, what may change, when development remains within the identity of that system, when a successor emerges, and whether earlier evaluations, alignment or safety claims remain attributable after transformation. This release contains six coordinated documents: the Canonical Reference, which presents the complete AI application of LA PROFILÉE; the AI Primer, which provides an accessible introduction; the Structural Map, which maps the complete architecture; the Application Handbook, which translates it into practical procedures; the Casebook, which demonstrates its application to characteristic AI transformations; and the Domain Guide, which connects the architecture to major fields and problems of artificial intelligence. Together, the six documents provide a canonical reference and a set of accessible routes for understanding, applying, testing, and challenging the application of LA PROFILÉE to artificial intelligence.

// Source

View paper (DOI)Open access versionOpenAlexZenodo (CERN European Organization for Nuclear Research)Published 2026-08-23

Authors: Marc Maibom