Same Robot, Different You: User Personality and Robot Role Effects in Large Language Model-Based Mobile Home Robot Interaction

This study investigates how robot role and user personality traits influence interactions with a large language model-based mobile home robot. Prior research has often examined these factors separately, focused on stationary robots, and relied on self-report measures that underrepresent actual conversational behavior. In a within-subjects design, 36 participants experienced assistant and companion roles with a mobile home robot. We examined subjective ratings and conversational metrics in relation to robot role and Big Five traits. The assistant role was rated higher on Animacy, Perceived Usefulness, and Intention to Use, whereas the companion role elicited longer mean turn lengths. A significant role × personality interaction showed that users low in Openness rated the companion robot lower on Perceived Usefulness. These findings suggest the value of adaptive role switching and personality-based personalization in the design of LLM-based mobile home robots.