Conversations for Vision: Remote Sighted Assistants Helping People with Visual Impairments
This research addresses accessibility challenges for people with visual impairments by analyzing real-world conversational support systems, though it is incremental in nature.
The study investigated how remote sighted assistants help people with visual impairments through conversational support, identifying four types of assistance and finding that communication styles adapt to context and involve leveraging personal knowledge and non-verbal cues.
People with visual impairment (PVI) must interact with a world they cannot see. Remote sighted assistance has emerged as a conversational/social support system. We interviewed participants who either provide or receive assistance via a conversational/social prosthetic called Aira (https://aira.io/). We identified four types of support provided: scene description, performance, social interaction, and navigation. We found that conversational style is context-dependent. Sighted assistants make intentional efforts to elicit PVI's personal knowledge and leverage it in the guidance they provide. PVI used non-verbal behaviors (e.g. hand gestures) as a parallel communication channel to provide feedback or guidance to sighted assistants. We also discuss implications for design.