Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Exploring Personality-Driven Personalization in XAI: Enhancing User Trust in Gameplay

Zhaoxin Li, Sophie Yang, Shijie Wang

TL;DR

The paper investigates how user personality influences preferences for XAI explanations and trust in interactive systems. It combines a Big Five personality survey (Mini-IPIP) with a simple MLP to predict preferences among decision-tree, textual, and factor-graph explanations, and tests this in a four-checkpoint navigation game with controlled XAI accuracy. Results show higher adherence to XAI suggestions and greater trust when the presented method aligns with the user's preferred XAI approach, supporting personalized XAI design. The study highlights practical implications for tailoring explainable interfaces to individual traits to enhance user engagement and confidence in AI systems.

Abstract

Tailoring XAI methods to individual needs is crucial for intuitive Human-AI interactions. While context and task goals are vital, factors like user personality traits could also influence method selection. Our study investigates using personality traits to predict user preferences among decision trees, texts, and factor graphs. We trained a Machine Learning model on responses to the Big Five personality test to predict preferences. Deploying these predicted preferences in a navigation game (n=6), we found users more receptive to personalized XAI recommendations, enhancing trust in the system. This underscores the significance of customization in XAI interfaces, impacting user engagement and confidence.

Exploring Personality-Driven Personalization in XAI: Enhancing User Trust in Gameplay

TL;DR

The paper investigates how user personality influences preferences for XAI explanations and trust in interactive systems. It combines a Big Five personality survey (Mini-IPIP) with a simple MLP to predict preferences among decision-tree, textual, and factor-graph explanations, and tests this in a four-checkpoint navigation game with controlled XAI accuracy. Results show higher adherence to XAI suggestions and greater trust when the presented method aligns with the user's preferred XAI approach, supporting personalized XAI design. The study highlights practical implications for tailoring explainable interfaces to individual traits to enhance user engagement and confidence in AI systems.

Abstract

Tailoring XAI methods to individual needs is crucial for intuitive Human-AI interactions. While context and task goals are vital, factors like user personality traits could also influence method selection. Our study investigates using personality traits to predict user preferences among decision trees, texts, and factor graphs. We trained a Machine Learning model on responses to the Big Five personality test to predict preferences. Deploying these predicted preferences in a navigation game (n=6), we found users more receptive to personalized XAI recommendations, enhancing trust in the system. This underscores the significance of customization in XAI interfaces, impacting user engagement and confidence.
Paper Structure (14 sections, 3 equations, 6 figures)

This paper contains 14 sections, 3 equations, 6 figures.

Table of Contents

  1. Introduction
  2. Experimental Design
  3. Survey Design
  4. The Big Five Personality
  5. XAI Preferences
  6. Machine Learning Model Design
  7. The Game
  8. User Study Design
  9. Metrics
  10. Results
  11. Willingness of following XAI's suggestion
  12. Trust in using the system
  13. Perceived usability
  14. Conclusion

Figures (6)

  • Figure 1: Capturing XAI preferences in Survey
  • Figure 2: Overview of the game design.
  • Figure 3: Three examples of XAI methods presented in the game (up: decision tree, middle: text, bottom: factor graph). The knowledge shared by both the agent and users is highlighted in orange in the decision tree and factor graph.
  • Figure 4: The percentage of users' decisions that align with XAI's instruction.
  • Figure 5: Users' trust in using the system.
  • ...and 1 more figures