← All reviews

The Guide Has Your Back: Exploring How Sighted Guides Can Enhance Accessibility in Social Virtual Reality for Blind and Low Vision People

Jazmin Collins, Crescentia Jung, Yeonju Jang, Danielle Montour, Andrea Stevenson Won, Shiri Azenkot · 2023 · Proceedings of the 25th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2023) · doi:10.1145/3597638.3608386

Summary

This paper presents the first exploration of using sighted guides to enhance accessibility in social virtual reality for blind and low vision (BLV) people. The researchers built a VR prototype in Unity deployed on Meta Quest 2 headsets, featuring three simulated park environments where a sighted guide could assist BLV users through navigation, object identification, and environmental description. Sixteen BLV participants (8 blind, 8 low vision, ages 29-61) completed three tasks: free exploration, a scavenger hunt, and free exploration with a flying guide. The guide followed behavioral guidelines emphasizing a passive, responsive approach rather than leading. The prototype included a novel "Shared Movement" feature inspired by physical sighted guidance, allowing BLV users to grab the guide's avatar arm to move together through the virtual space. Audio cues provided feedback for movement, surface materials, collisions, and snap turns. The study examined how participants interacted with the guide, what information they requested, how they communicated, and their preferences for guide initiative, functions, appearance, and visibility. The researchers observed granular participant-guide interactions and collected both quantitative ratings and qualitative interview data to understand the design space for virtual guidance systems.

Key findings

Participants rated the guide experience highly: usefulness 4.8/5, comfort 4.4/5, and communication effectiveness 4.4/5. Blind and low vision participants used the guide differently — blind participants relied more heavily on Shared Movement throughout their experience, while low vision participants experimented more with independent controls and used Shared Movement selectively for difficult visual tasks. Participants perceived the guide in three ways: as a utilitarian tool (90%+ requests being task-focused), as a companion they built social rapport with, or as a superfluous presence that hindered independence. Four participants felt the guide created unwanted dependence. Participants wanted varying levels of guide initiative — some preferred proactive alerts about obstacles, while others valued the passive approach. Importantly, participants wanted guides to modify inaccessible environments directly, such as adding sound beacons to objects or changing colors for low vision users. When asked about AI-powered guides, participants reacted positively about increased independence and availability, but raised concerns about trust, reliability, and loss of social interaction. Several participants suggested AI guides as primary tools with human backup.

Relevance

As virtual reality becomes increasingly mainstream, this research addresses a critical gap in making social VR spaces accessible to BLV users. The findings offer practical design guidance for VR developers: implement shared movement mechanics, provide customizable guide initiative levels, and consider both human and AI guide models. The distinction between blind and low vision interaction patterns is particularly valuable for designing adaptive assistance. The paper also raises important questions about dependence versus independence in assistive technology design — a tension relevant far beyond VR. For accessibility practitioners, the participants' desire for guides to modify inaccessible environments rather than just describe them represents a paradigm shift worth considering across all digital accessibility work. The research provides a foundation for future AI-powered accessibility tools in immersive environments.

Tags: virtual reality · social VR · blind and low vision · sighted guide · navigation · assistive technology · avatar accessibility · AI guide