Robots with Attitudes: Influence of LLM-Driven Robot Personalities on Motivation and Performance

TL;DR

The study investigates how LLM-driven robot personalities, particularly agreeableness, affect likability, intrinsic motivation, and cooperative performance in a human-robot Quickdraw task. Using an online pre-study and a lab-based main study with Vicuna-based personalities, it finds that agreeableness increases likability and perceived safety but does not consistently elevate motivation or performance across all participants. Correlations suggest that perceived openness and agreeableness of the robot can be linked to better task outcomes for some users, indicating potential benefits of personalized personality fitting. Overall, the work demonstrates the viability of using LLMs for stable robot personalities and highlights nuanced effects of personality on HRI outcomes, motivating further multi-trial and personalized investigations.

Abstract

Large language models enable unscripted conversations while maintaining a consistent personality. One desirable personality trait in cooperative partners, known to improve task performance, is agreeableness. To explore the impact of large language models on personality modeling for robots, as well as the effect of agreeable and non-agreeable personalities in cooperative tasks, we conduct a two-part study. This includes an online pre-study for personality validation and a lab-based main study to evaluate the effects on likability, motivation, and task performance. The results demonstrate that the robot's agreeableness significantly enhances its likability. No significant difference in intrinsic motivation was observed between the two personality types. However, the findings suggest that a robot exhibiting agreeableness and openness to new experiences can enhance task performance. This study highlights the advantages of employing large language models for customized modeling of robot personalities and provides evidence that a carefully chosen agreeable robot personality can positively influence human perceptions and lead to greater success in cooperative scenarios.

Figures (8)

• Figure 1: Main study between-subjects design with either an agreeable or non-agreeable robot in the cooperative Quickdraw task. The robot's personality is represented by a large language model that enables unscripted conversation and reactions to the participant's drawing.
• Figure 2: Part of the dialogue between a human and a chatbot used in the pre-study. The non-agreeable robot appears less interested in the task and is uncooperative.
• Figure 3: Interaction with an agreeable or non-agreeable robot in a cooperative drawing task. The participants need to sketch an object (in the image, the requested object is a pizza), while the robot comments and attempts to guess the correct object.
• Figure 4: Perception of the chatbots in the pre-study. Values indicate p value, $^{*}$p$<$ .05, $^{***}$p$<$ .001.
• Figure 5: Assessed personality traits of both chatbots in the pre-study using TIPI. Values indicate p value, $^{**}$p$<$ .01, $^{***}$p$<$ .001.