ECA — Endogenously-Constrained AI

This document defines Endogenously-Constrained AI (ECA) as the implementation layer of the Crowd-Based Dynamics framework.
It translates CBD structural limits into internal constraints governing artificial systems and socio-technical architectures.
ECA embeds constraints endogenously, avoiding ex post external control mechanisms.
The text grounds ECA in the CBD canonical formula, used as a design grammar rather than a predictive model.
It addresses temporal governability, saturation management, mimetic load, and reversibility in artificial agents.
Clear boundaries are set, excluding prediction, optimization, or normative enforcement.
The document positions ECA as a constrained implementation ensuring structural alignment under scale and complexity.