“Appropriations of Artificial Intelligence for Writing by Visually Impaired Students at University: Benefits, Barriers, and Risks”
Visual impairment; Artificial Intelligence; Writing; Inclusion; Higher Education.
This dissertation examines the different ways in which Generative Artificial Intelligence (GenAI) has been used by visually impaired students at the university, with the aim of critically understanding whether AI for writing has been used by these individuals in conjunction with other technologies. Furthermore, the research examines what visually impaired students say about the benefits, barriers, and potential risks in the context of using GenAI for academic writing. This dissertation works from the perspective of the Social Model of Disability. Semi structured interviews were conducted with eight visually impaired students from the University of Brasília (UnB), five undergraduates and three graduates. Two of these were blind, five had low vision, and one had monocular vision. Among the results obtained, I highlight "Time Optimization," "Information Comprehension," and "Text Revision" as the subcategories with the most context units identified as "Benefits." Regarding "Risks," the most common were "Plagiarism," "Inaccurate Information," and "Weakness in Training." The main subcategories related to "Barriers" were "Inaccessible Texts," "More Time for PwD," and "Cognitive Process." It can be concluded that visually impaired students at universities benefit from the use of AIgen for academic writing. However, the use of this technology is also fraught with risks and barriers, as these tools are not yet fully accessible to meet the specific needs of these users.