Understanding the Video Content Creation Journey of Creators with Sensory Impairment in Kenya
Lan Xiao, Maryam Bandukda, Franklin Mingzhe Li, Mark Colley, Catherine Holloway · 2025 · Proceedings of the 27th International ACM SIGACCESS Conference on Computers and Accessibility (ASSETS 2025) · doi:10.1145/3663547.3746356
Summary
This paper presents the first in-depth qualitative study of video content creation by people with visual and hearing impairments in a low- and middle-income country (LMIC) context, specifically Kenya. The researchers conducted semi-structured interviews with 20 video creators in Nairobi—10 with visual impairments (blindness and low vision) and 10 who are deaf or hard of hearing (DHH). Guided by the Interdependence Framework and the Ability-Diverse Collaboration Framework, the study examined how creators navigate video production workflows, the tools and platforms they use, the genres of content they produce, and the accessibility barriers they encounter across the entire production pipeline. Most participants relied primarily on smartphones for their entire workflow, using apps like TikTok, CapCut, and InShot, with some incorporating AI tools such as ChatGPT, Seeing AI, Google Lookout, and InVideo AI. The study identified video creation as a fundamentally staged and collaborative process involving five phases: pre-production, filming, visual editing, sound editing, and post-production. Collaboration was nearly ubiquitous, with 19 of 20 participants engaging human collaborators or AI tools at one or more stages. The research surfaces how creators actively negotiate agency, trust, and creative control while working with both human partners and AI technologies to bridge sensory gaps in resource-constrained environments.
Key findings
Creators faced challenges that were both universal to sensory-impaired content creators globally and specific to the LMIC context. BPS creators struggled with camera framing, inaccessible visual interfaces, and instability of accessibility features after software updates. DHH creators faced difficulties with audio editing, captioning accuracy, and producing content in sign language. These technical challenges were compounded by financial barriers, device limitations, absent training opportunities, language barriers (particularly serving Swahili-speaking audiences with English-centric tools), and platform monetization disadvantages that created "double digital divides." Collaboration patterns differed significantly between groups: DHH creators co-developed scripts with peers and relied on interpreters and hearing collaborators for audio-related tasks, while BPS creators typically delegated visual editing entirely but maintained strong creative direction through detailed verbal instructions. Six participants integrated AI tools directly into their workflows for scripting, captioning, scene description, and video generation. Creators described a model of "negotiated agency," fluidly shifting between director, collaborator, and editor roles based on the task, privacy preferences, and trustworthiness of available support. Collaboration was deeply relational, built on close-knit informal networks of friends (cited by 16 participants), family, and interpreters rather than formal professional support.
Relevance
This study significantly expands the geographical and sociotechnical scope of accessibility research in content creation by centring creators in a low-resource context rather than the high-income settings that dominate the literature. For practitioners and tool designers, the paper offers concrete design recommendations across four areas: supporting interdependent collaboration through multimodal interfaces and asynchronous distributed tools; supporting creator agency through explainable AI outputs and customisable automation; supporting authentic expression through multilingual captioning, adaptive non-visual editing templates, and expressive sign language interfaces; and addressing LMIC-specific infrastructure gaps through compatibility with affordable hardware, integrated digital literacy support, and community network strengthening. The concept of "double digital divides"—where disability-related accessibility barriers intersect with Global South infrastructural constraints—is particularly valuable for organisations working on global digital inclusion. The findings challenge the dominant framing of assistive technology as individualised support, demonstrating that accessibility often emerges through collaborative, interdependent practices.
Tags: video content creation · sensory impairment · visual impairment · deaf and hard of hearing · low- and middle-income countries · interdependence · ability-diverse collaboration · AI tools · digital inclusion · Global South