The Trusted Listener: The Influence of Anthropomorphic Eye Design of Social Robots on User's Perception of Trustworthiness

Title of the Paper

The Trusted Listener: The Influence of Anthropomorphic Eye Design of Social Robots on User’s Perception of Trustworthiness

Paper Information

• Subject Area: Human-Computer Interaction, Social Robots, Trust Perception
• Keywords: Human-Computer Interaction, Social Robots, Anthropomorphic Design, Robot Eye Design, User Trust Perception

Research Background and Problem

• Problem or Challenge: With the development of artificial intelligence and human-computer interaction technologies, social robots are widely applied in scenarios such as healthcare, education, and households. Trust is a key factor in achieving a high-quality human-robot interaction experience. However, there is currently a lack of in-depth research on the design of robot eyes in a "listening" state and its impact on users' trust perception.
• Significance: Visual characteristics of robots, particularly eye design, are critical in forming users' first impressions during interactions, directly influencing their level of trust in the robot.
• Research Motivation and Related Work: Numerous studies have shown that robots' voice, language style, and non-verbal signals (e.g., gaze) positively influence user experience and trust perception. However, there is insufficient research on the dynamic design of robot eyes in a "listening" state (e.g., blink rate, gaze aversion).

Solution

• Proposed Method or Solution:

  • This study proposes an experimental approach to investigate the impact of social robots' eye design (visual complexity, blink rate, gaze aversion behavior) on users' trust perception.
  • Research hypotheses include: 1) Higher visual complexity in eye design leads to higher trust in the robot; 2) Within a reasonable range, a lower blink rate helps establish trust; 3) Robots with gaze aversion in a "listening" state are perceived as more trustworthy.

• Key Technologies and Implementation Steps:

  • Development of a custom physical robot model, "eBot," with an embedded screen to display various eye animations.
  • Use of three experimental variables:
    1. Visual Complexity: Design of eye animations at three levels of visual complexity (low, medium, high).
    2. Blink Rate: Setting two rates (10 blinks/minute and 20 blinks/minute).
    3. Gaze Aversion: Simulation of behavior with/without gaze aversion.
  • Data collection from 66 university student participants through Wizard of Oz experiments, using a trust perception scale to measure user experience.

Research Findings

• Specific Findings:

  • Visual Complexity: Higher visual complexity in eye design significantly increases users' trust in the robot, particularly in the benevolence dimension.
  • Blink Rate: Blink rate did not show a significant impact on trust perception, but a slightly higher-than-human average blink rate was considered more comfortable in interaction experiences.
  • Gaze Aversion: Robots with gaze aversion behavior were perceived as significantly more trustworthy than those without such behavior.

• Advantages Compared to Existing Solutions:

  1. The study examines the impact of multi-dimensional variables in robot eye design in dynamic scenarios, enhancing practical application value.
  2. Provides actionable design recommendations, such as adjustable eye characteristics and dynamic behavior settings.

• Experimental or Evaluation Results:

  • Robots with high visual complexity scored significantly higher than those with medium or low complexity.
  • Robots with gaze aversion behavior significantly improved user trust scores in the dimensions of competence, benevolence, and integrity.
  • Although blink rate showed no significant impact, users rated robots with lower blink rates as more reliable.

• Limitations and Future Directions:

  1. Sample Size Limitation: The study involved only 66 students, with a limited and non-diverse sample size. Future research should expand participant diversity.
  2. Lack of Interaction Effect Studies: The study examined single variables independently without exploring the combined effects of multiple variables on trust perception.
  3. Other Design Features: Future research could integrate more non-verbal features, such as dynamic facial expressions and head movements, to further enhance trust perception studies.
  4. Practical Performance: While perceived trust is important, it should be combined with functional performance (e.g., voice response capabilities) to enhance trust in practical applications.

Summary and Contributions

• Research Significance:

  1. Addresses the research gap on the impact of social robots' eye design on users' trust perception.
  2. Proposes experimental-based preliminary design guidelines, providing design references for enhancing trust in social robots.

• Research Contributions:

  • Provides experimental evidence that visual complexity and gaze aversion behavior significantly influence user trust perception.
  • Highlights the importance of personalized and adjustable animation parameter settings in future social robot design.