Generative User Interfaces and UI Design Practice: A Socio-Technical Study of Implications and Adoption Requirements

Topic
This master thesis investigates how Generative User Interfaces affect user interface design practice from a socio-technical perspective. GenUI refers to interfaces that are created, adapted or refined through generative AI processes rather than being fully predefined before deployment. The thesis examines how this shift changes design workflows, professional roles, collaboration, risks, ethical responsibilities and the technical as well as organisational requirements needed for effective and responsible GenUI adoption.

Relevance
GenUI is relevant for practitioners because it may fundamentally change how digital interfaces are designed, evaluated and governed. Instead of designing fixed screens, UX/UI professionals may increasingly define rules, constraints, design systems and generative spaces. This creates opportunities for faster ideation, prototyping and personalisation, but also introduces challenges around consistency, accessibility, accountability, data quality, user control and organisational governance. Practitioners therefore need guidance on how to adopt GenUI responsibly.

Results
The thesis shows that GenUI shifts UI design from producing static interface artefacts toward specifying, orchestrating and validating dynamic interface systems. Designers are not replaced, but their role becomes more strategic, technical and coordinative. The main UX tension is between personalisation and consistency: GenUI can create more contextual interfaces, but may also reduce predictability, orientation and trust. Responsible adoption requires modular design systems, reliable data, traceability, fallbacks, human validation, governance structures and clear accountability.

Implications for practitioners

• Treat design systems as critical infrastructure: components, design tokens, accessibility rules, brand principles and interaction patterns must guide what GenUI systems are allowed to generate.
• Shift quality assurance from testing fixed screens to testing generative behaviour across different users, prompts, contexts and edge cases.
• Define governance before deployment, including ownership, accountability, data use, compliance boundaries, incident response and escalation paths.
• Start with bounded, low-risk use cases before applying GenUI to legally binding, financial, medical or system-critical processes.
• Preserve user control through standard views, fallback options, transparency and mechanisms that allow users to stop, revise or recover generated interface states.

Methods
The thesis follows a Design Science Research approach based on Hevner’s Three-Cycle View, connecting practical relevance, theoretical grounding and conceptual artifact development. First, a systematic literature review was conducted to establish the theoretical basis on Generative AI, GenUI, UI design and Socio-Technical Systems Theory. Second, 14 semi-structured expert interviews were conducted with UX and UI professionals, of which 13 were included in the final analysis. The interview data were analysed using the Gioia methodology, moving from first-order codes to second-order themes and aggregate dimensions. This enabled the development of empirically grounded implications and requirements for GenUI adoption.