"When We're Looking at the Robot, We See Each Other": A Comparison of Robotic, Mirror-Based, and Hybrid Interventions for Stranger Interaction

Paper Title

"When We're Looking at the Robot, We See Each Other": A Comparison of Robotic, Mirror-Based, and Hybrid Interventions for Stranger Interaction

Publication Info

• Topic area: Human–Robot Interaction (HRI) and Human–Computer Interaction (HCI) for fostering social connections.
• Keywords: Stranger interaction, social robots, mediated gaze, adaptive mirrors, public spaces, eye contact, serendipitous encounters, socio-spatial interfaces, ice-breaking technologies, interpersonal closeness.

Background and Problem

• Problem / challenge: Public spaces often suppress social encounters due to “civil inattention,” compounded by the ambiguity and risk of initiating contact with strangers. Existing technologies for mediated gaze and social signaling lack systematic understanding of how mobile, reflective objects can facilitate stranger interactions.
• Significance: Enhancing stranger interactions can improve happiness, belonging, and social cohesion, addressing a meaningful challenge for HCI and HRI.
• Motivation and related work: Prior work has explored interactive furniture, gaze-aware robots, and playful cues to nudge stranger interactions, but has not systematically examined mobile mirrors as mediators of gaze and social signaling.

Solution

• Proposed approach: MirrorBot, a mobile robot equipped with adaptive mirrors, designed to choreograph mediated eye contact and shared attention between strangers.
• Novelty:
  1. Introduction of MirrorBot, leveraging mirrors as symbolic and interactive artefacts for social connection.
  2. Empirical comparison of robot movement and mirrors in facilitating stranger interactions.
  3. Design implications for ice-breaking technologies using mirrors as a social material.
• Procedure and key techniques:
  • MirrorBot uses adaptive mirrors with four key states: Voids, Dual Self, Dual Other, and Self + Other, to gradually foster mutual awareness and eye contact.
  • A controlled 2×2 experiment tested four conditions (MirrorBot, Bot-only, Mirror-only, Control) with 90 participants (45 dyads) to evaluate conversation initiation, interpersonal closeness, togetherness, and interaction quality.
  • Behavioral coding and validated survey scales measured outcomes, supported by qualitative interviews.

Results

• Concrete findings:
  • MirrorBot participants reported the highest ease and lowest awkwardness in initiating conversations (83% initiated vs. 39% in Bot-only).
  • Perceived interpersonal closeness was highest in MirrorBot (shortest self–other distance: M = 124.85 pixels).
  • Behavioral measures showed significantly higher shared attention (23% session time) and emotional revealing (8.8%) in MirrorBot compared to other conditions.
  • Positive affective experience and enjoyment were significantly higher in MirrorBot compared to Mirror-only.
• Advantage over baselines:
  • MirrorBot outperformed static mirrors and non-reflective robots in fostering mutual awareness, reducing awkwardness, and enhancing interaction quality.
  • Static mirrors (Mirror-only) often amplified self-consciousness and awkwardness, while Bot-only robots competed for attention rather than facilitating human–human interaction.
• Experiments / evaluation:
  • Participants were randomly assigned to one of four conditions in a waiting-room scenario under a memory task cover story.
  • Mixed-effects models analyzed survey data, while video recordings provided behavioral metrics.
• Limitations and future work:
  • Limited ecological validity due to controlled lab setting; field studies in diverse public spaces are needed.
  • Lack of clean separation between mirror motion and robot embodiment effects.
  • Teleoperated MirrorBot limits autonomy; future work should explore adaptive policies for timing and engagement.
  • Individual differences in personality traits require larger sample sizes for systematic analysis.

Summary

MirrorBot demonstrates the potential of mobile robotic mirrors to facilitate stranger interactions by choreographing mediated gaze and shared attention. Empirical findings show that MirrorBot reduces barriers to initiating conversations, fosters interpersonal closeness, and enhances interaction quality compared to static mirrors and non-reflective robots. While the study highlights the importance of timing, framing, and subtlety in designing ice-breaking technologies, future work should address autonomy, ecological validity, and individual differences to refine and expand the system’s applicability in real-world settings.