Web Application Accessibility with Artificial Intelligence

Web accessibility is a hot topic driven by social inclusion that has brought the requirement of making content accessible to all users. It is estimated that one billion people worldwide have various disabilities that require some kind of assistive technology, and this number only increases as the population ages. Content authors and web developers are faced with increasingly more work and responsibilities to keep up with the requirements for making the web accessible to all. Artificial intelligence is a catch-all name for a group of technologies that imitate some aspect of human intelligence and are used in a variety of applications ranging from healthcare to digital marketing. The purpose of this thesis was to examine whether artificial intelligence can be leveraged for tasks that could help improve web application accessibility. There are many useful applications of artificial intelligence within this field and the number of potential applications only increases as this technology evolves. One such task that this thesis examined was the use of automatic image captioning. Captions are important for image accessibility but adding descriptions for images is a time-consuming task that is often overlooked. The use of automatic image captioning could however facilitate and streamline the task of making images accessible to all. This thesis uncovers how a modern web application can be built to integrate with this kind of machine learning technology. The literature review of this work is made up of three chapters that present the concept of web accessibility and its challenges, how artificial intelligence can improve web accessibility, as well as the theory of machine learning methods such as neural networks. These chapters lay the foundation for understanding the practical work of this thesis and why the effort within this field is necessary