User experience design in artificial intelligence applications

Ivana Števulová

Supervisor(s): Ing. Martin Dubovský

Slovak Technical University

Abstract: Artificial intelligence (AI) is increasingly used in medical diagnostics, yet many AI systems remain difficult for clini- cians to understand, creating challenges in trust, safety and adoption. This work explores how explainability meth- ods can be integrated into the user interface of AI-assisted medical tools through an intelligent assistant implemented as a conversational agent for domain experts. The design is situated in the diagnosis of Primary Cil- iary Dyskinesia (PCD), a rare genetic disorder diagnosed through Transmission Electron Microscopy (TEM). Drawing on human-centered design principles and the EUCA (End-User-Centered Explainable AI) framework, the prototype integrates explanation methods such as un- certainty maps, concept alignment, rule-based explanation and more, which are described in section 2 into a con- versational interface embedded in the CiliaScan medical analysis tool. Prototypes were developed through iterative low and high-fidelity design with user scenarios grounding the interface in realistic clinical workflow. The aim of this work is to design and prototype explain- ability and interpretability methods in AI-infused user in- terface in the medical domain, focusing on how AI sys- tems can present the reasoning behind their predictions and responses to support domain experts understanding and trust.
Keywords: Design, Human-Computer Interaction
Full text:
Year: 2026