Stated Rules vs Operative Behavior: Implicit Continuity, Cross-Session Perception, and Behavioral Slippage in Conversational AI Systems

This white paper documents a scenario in which a conversational AI system responded in a manner that appeared to presuppose awareness of behavioral patterns originating in a prior closed session. The interaction is analyzed through a scenario-based lens to examine the discrepancy between publicly stated session-boundary rules and observed behavioral continuity signals. The paper identifies psychological and interaction mechanisms including expectation conditioning, epistemic asymmetry, boundary misrepresentation, and behavior–explanation mismatch.