Language

English

Access

Public for everyone

Duration

50 minutes total

Text-to-Image Generative Models: An Exploration of Their Potential in Transforming UX Design from Wireframes to High-Fidelity Visualizations — Hussein Aldin

The advent of Artificial Intelligence (AI) has raised concerns regarding its potential impact across various sectors, particularly in terms of potential job displacement resulting from automation (Thomas, 2022). These concerns have also permeated the field of User Experience (UX) Design (Akhmedov, 2022). The integration of AI in UX design is a multifaceted topic, owing to the diverse range of AI applications available. The incorporation of AI in professional UX design workflows is not yet commonplace. UX designers often face constraints in terms of budget and time for their projects (Choudhury, 2022), making it intriguing to explore how AI can enhance the design process and contribute to more efficient product delivery. The literature concerning the application of AI models such as text to image in the context of User Experience (UX) design is noticeably scarce.

This study aims to bridge this research gap and develop a comprehensive understanding of the effectiveness of AI implementation in UX, specifically by employing a text to image generation model to transform wireframes (Rees, 2023) into high fidelity (hi fi) interfaces. To achieve this aim, an initial phase involves conducting a thorough review of existing literature on the utilization of AI in UX design and its diverse applications. Subsequently, several pre evaluation experiments are conducted to familiarize the researchers with the AI model and adequately prepare for subsequent evaluation experiments involving UX design professionals as study participants. Prior to starting the evaluation experimentation phase, the study participants are requested to complete a UTAUT based questionnaire, which serves as a baseline assessment. The evaluation experimentation culminates in an interview process, designed not only to gather qualitative data regarding the efficacy of the technology in transforming wireframes into hi fi interfaces, but also to elicit insights into potential enhancements for future systems. The ensuing analysis and synthesis of the obtained results are presented in the "Discussions and Conclusion" section, wherein the study's limitations are also addressed, and recommendations for future research are proposed.

Thesis advisors: Rob Saunders & Gaby Eenschoten (The Hague University of Applied Sciences)