Bachelor of Science in Software Engineering: Generative AI’s impact on user interfaces in data-intensive web applications: An exploratory analysis
As technology advances, more and more companies have becomeinterested in integrating generative AI into their web applications. Generative AI isan artificial intelligence that generates new data that mimics original datasets andhas the potential to transform user interfaces by making them more intuitive andimproving the user experience. Objective: The objective of this thesis is to investigate whether generative AI canimprove the user experience in a data-intensive application with a focus on usabilityin terms of learnability and understandability. The work also focuses on investigat-ing how one’s perceived confidence and efficiency is when performing tasks with andwithout generative AI in a data-intensive application. Another focus area is to assessthe quality and relevance of the information generated by generative AI. Methods: In this study, an experiment was conducted by integrating a custom fine-tuned generative pretrained transformer (GPT) based on GPT-4 with a traditionalgraphical user interface in a data-intensive web application. The user experience andusability of the AI augmented interface was compared to a traditional interface us-ing a mixed method consisting of A/B testing, usability testing, and semi-structuredinterviews. Results: The results from the experiment show that there were mixed results forthe participants who had access to the traditional interface enhanced with generativeAI. They experienced varying degrees of improved efficiency and engagement. How-ever, challenges related to AI response times and reliability had a negative impact onoverall effectiveness. This resulted in the application augmented with generative AInot outperforming the traditional interface approach in the usability measurements. Conclusions: We observed that there is potential for generative AI to improvethe user experience in data-intensive applications, but it is highly dependent on theeffectiveness of generative AI’s accuracy and responsiveness. Further research, de-velopment, and optimization is required for generative AI to have the potential toimprove usability and user experience.