Accessible Digital Content Creation by Blind and Low Vision Individuals

Abstract

Creative activities are central to cultural participation, self-expression, and professional life, yet they remain largely inaccessible to blind and low vision (BLV) individuals. Most creative tools assume a sighted creator by default, offering limited nonvisual feedback to support iterative decision-making with visual content. Meanwhile, accessibility research has historically positioned BLV individuals as consumers of digital media rather than creators, leaving critical gaps in our understanding of their creative needs, goals, and workflows. This dissertation examines how BLV individuals engage in visually mediated creative work and how AI-assisted tools can better support their creative needs. Through a user-centered, iterative approach spanning four studies, it identifies the varied feedback demands that arise across creative contexts and explores how those demands can be addressed through accessible design. First, a large-scale formative study maps the landscape of BLV creative practices, revealing strong and varied interests in visual content creation that challenge common assumptions about what BLV individuals want to make, while surfacing the distinct, task-specific feedback demands these interests entail. Then, two design explorations examine AI-powered creative support for distinct creative contexts that the formative study identified as critical yet underserved: visual privacy management in photography, and expressive aesthetic configuration in visual media. Finally, a synthesis study distills visual feedback design insights for creative tasks into a framework via a systematic literature review and expert evaluation. Together, these studies surface how BLV individuals navigate systemic exclusion and technical barriers in creative work, and demonstrate that supporting creative agency requires tools that are responsive to users' goals, rich in contextual feedback, and grounded in non-visual ways of knowing and expressing. This dissertation contributes empirical understanding of BLV creative practices, design insights for accessible AI-assisted tool development, and a practical framework that equips researchers and practitioners to support BLV individuals' full authorship across visually mediated creative contexts.