Adaptive interface system for inclusive digital experiences

Abstract

Most websites are still designed to look and work the same for every user, which can leave out people whose needs or preferences do not match the defaults. This thesis looks at adaptive interfaces as a way to respond to that problem. The aim was to design and test a system that adjusts how a website looks and behaves based on what the user needs, while still keeping the user in control of any changes that happen. The research combined a literature review of 16 papers on adaptive interfaces and accessibility, an online survey with 56 respondents, and moderated user testing with three participants who had different accessibility needs. The literature covered areas such as cognitive load, dyslexia, ADHD, colour vision deficiency, and the effects of stress. The survey looked at how people currently experience accessibility on websites and what kind of control they would want over an adaptive system. The user testing was used to check whether the design decisions actually made sense to real users. The findings shaped a prototype built as a desktop browser extension, since the survey showed desktop was the most used device for longer browsing tasks. The system uses a two-tier structure with four main modes, which are Standard View, Vision Support Mode, Reading Enhancement Mode and Focus Mode, and two overlay settings that can be added on top of any mode. User testing showed that Vision Support Mode and Focus Mode were the most appreciated, but it also challenged some assumptions. Features like the OpenDyslexic font and syllable highlighting were found to be distracting rather than helpful, so they were moved to opt-in options. The thesis concludes that adaptive interfaces work best when they suggest changes rather than forcing them, and when users can always see, undo, or turn off what the system is doing. Clutter removal came up repeatedly in both the survey and the testing, and should be treated as a core feature in future versions rather than an optional extra. Future work includes mobile support, larger-scale testing in real browsing environments, and exploring how generative AI could be used for more personalised adaptations while keeping user data local.