Improving usability of library websites for visually impaired users by using artificial intelligence capabilities

The purpose of this research is to deal with some artificial intelligence solutions to help improving the usability of library websites for visually impaired users.

The current research is a qualitative that was conducted with the interview and think aloud protocol. In this research, through interview with visually impaired users, the best and worst features of library websites were questioned. Users were observed while using a library website and performing daily tasks while verbally expressing their thoughts, feelings, and opinions about their interaction experience. The interview was conducted individually and online or at a specific location. 4 large library websites of the country (the National Library and Archives of Iran; Organization of Libraries, Museums and Documents Center of Astan Quds Razavi; Library, Museum and Document Center Islamic Consultative Assembly and the Central Library and Document Center of Tehran University) were selected to perform the tasks. The research community was visually impaired users who were selected using the targeted sampling method in the number of 33 people. The analysis of the users' conversations was done based on qualitative content analysis, and the MaxQDA software was used and placed in 90 final codes and 3 general categories and 8 subcategories. These codes were continuously expanded and revised while reviewing the transcripts. Another researcher participated in the content analysis and reviewed the transcripts and extracted categories, and the data were evaluated several times.

The best features of the websites from the users' point of view were: standard and accessibility of the elements on the pages, valuable content and honesty in introducing and presenting the content, logical segmentation and headings and proper organization of page elements, allocation of Alt tag for graphics, optimal organization of results search and... the worst features of the websites were: image and unknown security codes (captchas), lack of automatic correction of keywords spelling mistakes, problems of online conversations and unknown sending and receiving messages, dynamic content, lack of adjustment of elements on the page with the keyboard, not having labels for graphics and user input elements, and non-principled page design, etc.

Automatic correction, intelligent voice assistants, result clustering, intelligent filtering, intelligent question and answer, text and image processing/image description, text summarization, semantic search and natural language and user interface personalization are some of the ways to improve the usability of library websites.