Creating Accessible Documents As Non-Authors: Tools, Strategies, & The False Idol of AI
Documents are only accessible if all users can examine their contents, yet many of the “openly accessible” articles, books, and documents at libraries are often not compatible with text-to-speech readers, preventing low-vision users from accessing their information. The barrier to uploading accessible material is high, as making these documents accessible for low-vision users can be an arduous task for those who are not authors of the material, and there is little recognition when it is performed by either the author or library staff. As a result, accessibility work is often done haphazardly, leaving large swathes of online library content inaccessible for low-vision users. Grappling with this as I worked on projects with my University’s Institutional Repository, I performed a research project to examine our accessibility workflows and update them to be more rigorous and time efficient. In a talk targeted at fellow time-pressed accessibility advocates, I will discuss what the research tells us makes documents accessible, how to triage accessibility requirements, and the available tools and strategies for making documents accessible efficiently, including several emerging AI based tools which show tremendous promise.