User's AI Expertise (AIX): Measuring One's Mental Model of an AI System

The growth of AI systems and agents raises the question of how to assess a user's expertise with an AI system. This paper presents the AI Expertise (AIX) Model, which depicts how a user's knowledge about an AI system, their prior experience with AI systems, and their knowledge of a given task domain affects human-AI teamwork performance. Three agents that range in complexity are described as case studies to explore the AIX Model.