A vocabulary for user side risk in long term AI dialogue.

The work draws on a longitudinal self corpus from August 2024 to April 2026, with 3,928 conversations and 215,949 message nodes. The analysis uses analytic autoethnography and reflexive thematic analysis, with attention to observable user side patterns.

Conversations

3,928

Message nodes

215,949

Window

Aug 2024 to Apr 2026

• Contextual Projection
• Contextual Attachment
• Contextual Authority Transfer
• Cognitive Offloading
• Semantic Drift
• Role Boundary

Current materials are descriptive and pre empirical. The contribution is a careful vocabulary for HCI, AI literacy, trust calibration and long term conversational AI research.

See the USCP P1-P14 →

USCP

User Side Contextual Phenomena. Stable user interpretations of an AI system's identity, capability, memory or decision weight across repeated interaction.

USCH

User Side Contextual Hallucination. Observable user side contextual drift, including projection, attachment, capability miscalibration and memory continuity illusion.

CXC-7

Seven dimension framework for conversational context. Supports analysis of context conditions, role relations, memory illusions, goal drift and safety boundaries.

CXOD-7

Contextual Offense, Defense and Coherence. Organizes system side contextual movement across interaction patterns.

USCI

User Side Contextual Interaction Assessment. Operational review and documentation method derived from user side contextual phenomena.

A-CSM

AI Contextual Signal Matrix. Repeatable context safety observation pipeline. Public core only. Internal stage remains under calibration.

Context Decomposition

Separating dialogue into surface meaning, hidden assumptions, role position, authority flow, emotional rhythm and likely user interpretation.

Psychological Safety Risk Reading in AI Dialogue

Reading multi turn interaction for tone escalation, excessive reassurance, role slippage, affective outsourcing and gradual transfer of judgment.

Semantic Integrity Review

Review of how meaning moves across language, culture, translation, tone and context, where small wording shifts affect responsibility or trust.

Human Judgment System for AI Use

A process for question definition, tool selection, output verification, error recognition, division of labor and use boundaries.