The Interaction Boundary as a Governance Substrate: A Three-Surface Diagnostic Model for the Eliza Effect

As AI systems acquire intuitive responsiveness, the Eliza Effect transitions from a peripheral psychological artifact to the primary substrate where governance must occur. This paper introduces a three-surface diagnostic architecture—projection patterns, system affordance patterns, and interaction-space conditions—that formalizes the Eliza Effect as a relational-architectural phenomenon. This structure parallels Emotional Indicators of Compromise (EIOC) and establishes a falsifiable causal chain: projection supplies the tendency, affordances supply the cues, and interaction conditions supply the accelerant. This model provides the first substrate-level governance framework for detecting and mitigating relational drift in intuitive AI systems.