Autistic Adults’ Perspectives on Self-Expression and Dyadic vs. Group Interactions in AR Telepresence

Paper Title

Autistic Adults’ Perspectives on Self-Expression and Dyadic vs. Group Interactions in AR Telepresence

Publication Info

• Topic area: Augmented Reality (AR) telepresence for social interaction among autistic adults
• Keywords: Autism, AR telepresence, avatar customization, social presence, dyadic interaction, group interaction, participatory design, sensory sensitivities, cognitive load, self-expression

Background and Problem

• Problem / challenge: Existing communication technologies for autistic individuals often fail to balance sensory control with social presence. Little is known about how autistic adults prefer to be represented as avatars or how they experience social presence in AR telepresence, particularly in dyadic vs. group interactions.
• Significance: Social isolation among autistic adults is linked to depression, and AR telepresence could provide a controlled, inclusive medium for fostering social connection while addressing sensory sensitivities.
• Motivation and related work: Previous studies have explored AR and VR for autistic children and adults but have rarely involved autistic adults in co-design processes. Research on avatar representation highlights the importance of self-expression, but its application to autistic adults remains underexplored. This paper addresses these gaps by focusing on avatar co-creation and social presence in AR-mediated dyadic and group interactions.

Solution

• Proposed approach: A participatory study involving autistic adults to co-create avatars and evaluate social presence in dyadic and group AR telepresence activities.
• Novelty:
  1. Insights into autistic adults’ preferences for self-representation through avatars.
  2. Comparative evaluation of social presence in dyadic vs. group AR interactions.
  3. Development of AR telepresence activities tailored to autistic participants’ needs.
  4. Design recommendations for neurodiversity-affirming AR systems.
• Procedure and key techniques:
  1. Phase 1: Avatar co-creation using Metahuman Creator, guided by participant feedback.
  2. Phase 2: Participatory design of AR activities (Checkers for dyadic, Apples to Apples for group) and environmental features based on sensory and social preferences.
  3. Phase 3: Evaluation of social presence using the Networked Minds Social Presence Inventory and post-activity interviews.

Results

• Concrete findings:
  • Dyadic interactions (Checkers) yielded higher social presence scores across all subscales (e.g., co-presence: 3.05 vs. 1.95 for group; p < 0.001).
  • Group interactions (Apples to Apples) introduced higher cognitive and perceptual load, reducing participants’ sense of connection.
  • Participants prioritized self-expression over physical realism in avatar design, with preferences influenced by social context.
• Advantage over baselines:
  • AR telepresence provided a higher sense of presence than traditional 2D video calls and was preferred over in-person interaction in scenarios involving sensory or logistical barriers.
• Experiments / evaluation:
  • Eight autistic adults participated in a three-phase study involving avatar co-creation, participatory design of AR activities, and evaluation of dyadic and group interactions.
  • Social presence was measured using a validated inventory, and qualitative insights were gathered through thematic analysis of interviews.
• Limitations and future work:
  • Small, homogeneous sample (n=8) limits generalizability.
  • Differences in activities (Checkers vs. Apples to Apples) may have influenced results.
  • Future work should explore personalization at scale, adaptive systems, and longitudinal studies to assess evolving comfort with AR.

Summary

This study explored AR telepresence as a medium for social interaction among autistic adults, focusing on avatar co-creation and social presence in dyadic and group contexts. Participants emphasized self-expression over physical realism in avatar design, with preferences shaped by identity and social context. Dyadic interactions fostered stronger social presence, while group interactions introduced cognitive and perceptual challenges. The findings highlight the potential of AR telepresence to support neurodiversity-affirming communication, with design recommendations for customizable avatars, reduced cognitive load, and improved motion fidelity. These insights pave the way for inclusive AR systems that enhance social engagement for autistic individuals.