Generative AI in UI design

Artificial Intelligence has made its mark in almost every single technological field, and User Interface design is no exception, raising the question of whether Generative AI could and should be used in order to augment or replace the design processes of human designers. This research study aims to explore the role of Generative AI in UI design and provide valuable insights to inform the decision-making process of designers, researchers, and stakeholders. A cross-sectional survey in the form of interviews was conducted with UI designers, graphic designers, and UX professionals from a prominent technology consulting company. The study focused on investigating their perceptions and potential utilisation of Generative AI tools across three dimensions: inspiration, collaboration or full automation (replacement). The designers were presented with AI-derived designs for evaluation, yielding varied feedback across different designs. The interviewees recognised the potential value of AI-generated designs in terms of time-saving, generation of compelling colour schemes, accessibility, and alignment with existing UI conventions. These findings contribute to the development of ethical and responsible AI practices in Human-Computer Interaction and provide insights for designing future digital goods and services. By delving into the intersection of Generative AI and UI design, this study not only addresses the research problem at hand but also empowers designers and stakeholders to make informed decisions regarding the integration of AI in their workflow. The results highlight the potential benefits of Generative AI tools while acknowledging the need for careful evaluation and consideration of their output. Ultimately, this research serves as a valuable resource for advancing the field of UI design, fostering responsible AI practices, and shaping the future of digital design.

In 2023, generative AI tools like Midjourney and Uizard went from novelty to mainstream almost overnight. They could produce a full app screen in under a minute. That raised a question the design industry hadn't answered: is GenAI a tool that makes designers better, or one that makes them unnecessary? Most of the debate ran on hype and fear. Very little of it came from the professionals actually doing the work.

For my BSc thesis at the Athens University of Economics and Business, I designed and ran a structured interview study with 7 design professionals at a major technology consulting company. They were UI, UX, graphic and product designers, with experience ranging from under a year to over a decade.

• I generated 5 AI-made interfaces (a book app, travel app, food delivery app, car rental site and e-commerce site) using Midjourney and Uizard. Each took between 28 and 55 seconds.
• Participants rated each design on four core UI principles: consistency, clarity, accessibility and visual appeal.
• I framed the interviews around three possible roles for GenAI: inspiration, collaboration, or replacement. I captured both quantitative ratings and open-ended views, including ethical concerns.

The verdict leaned clearly toward augmentation, not replacement. Only one design was rated "Good", one "Fair", and three "Average". Designers praised the speed, the colour schemes, the accessibility, and how closely outputs followed established UI conventions. They consistently flagged weak UX logic, unclear information flow, and a tendency toward generic, "default" results. Writing the right prompt also turned out to be a skill of its own. Most saw GenAI as a starting point for ideas, and several actively wanted it to take over repetitive execution work. None believed it could replace design thinking, empathy, or an understanding of business context. The study captured how practitioners saw these tools at an early turning point, and it set out what AI would need to improve before it could compete with human designers.