Designing Space for Human Meaning in an Automated World
Yunxing Liu鈥檚 doctoral research shows how hybrid interfaces and carefully applied AI can scale qualitative insight while preserving people鈥檚 own perspectives in design, entrepreneurship, and societal decision-making.
has earned his doctorate at Eindhoven University of Technology with research on how people can better express what truly matters to them through AI-supported research tools. On May 21, 2026, Liu defended his dissertation within the Department of Industrial Design and EAISI. His work focuses on careful automation: using AI and hybrid interfaces to support research at scale while keeping human perspective and researcher responsibility at the center.
Understanding the Human Why
Designers, entrepreneurs and policymakers increasingly rely on data to guide decisions, yet many challenges they face are not only about what people choose, but also why those choices matter. Whether a start-up is testing a new service, a hospital is evaluating patient experiences, or a city is shaping mobility policy, understanding personal values and expectations is essential. Traditional interviews can reveal this depth, but they are slow and hard to scale, while surveys often flatten answers into predefined boxes.
Hidden Distinctions
Liu鈥檚 research builds on the Repertory Grid Technique, a qualitative method that asks people to compare examples and describe similarities and differences in their own words. These distinctions reveal how individuals make sense of products, technologies or situations. For entrepreneurs, this can surface why one platform feels trustworthy while another does not. For societal applications, it can explain how people perceive digital services, healthcare tools or educational technologies beyond surface-level satisfaction scores.
Hybrid Interface Design
Within the Designing Quality in Interaction research group, Liu designed and evaluated Q-Survey, a hybrid research interface that combines conversational guidance with visual structure. Participants are guided step by step through comparisons, while visual elements keep complex tasks understandable and reviewable. Later versions explored how large language models can help participants clarify and reflect on their answers, while allowing participants to remain in control of their own wording.
Careful Automation
The research shows that AI can support qualitative research at scale when used with restraint. Automation works best when it handles repetitive and procedural tasks, such as presenting comparisons and recording responses, while researchers remain responsible for study design and interpretation. This human-in-the-loop approach helps entrepreneurs and researchers test ideas with richer feedback and allows organizations to learn from diverse voices without losing nuance.
Societal Value
By showing how AI-supported interfaces can amplify rather than replace human expression, Liu鈥檚 work contributes to more responsible uses of AI in research and design. It offers practical guidance for building tools that respect accountability and transparency, which is increasingly important as AI systems shape everyday decisions.
Yunxing Liu defended his thesis on May 21, 2026.
Title of the thesis: .
Supervisors: Prof. dr. ir. Jean-Bernard Martens and Prof. dr. Panos Markopoulos.