The following papers will be presented during the workshop:

Session A: Methods and Frameworks for Usability

Explorative User Research Methods in Medical Applications. Erika Vaczlavova, Miroslav Laco and Wanda Benesova (Slovak University of Technology, Bratislava, Slovakia)

In every phase of the user-centred design (UCD) process, the focus is on the essential methods of bringing users to the centre - the user research. An appropriate user research strategy and correctly chosen research methods can make the design process more efficient and enhance the communication between domains. The approach to user research strategies and methods is even more critical for applied UCD, in collaboration within specialised domains, such as medicine. The domain experts from the field of medicine must obtain trust in the collaboration within the design process and in the designed system itself. Moreover, their resources and involvement must be considered wisely. This can be achieved by effective knowledge transfer methods and tailoring research methods, such as exploratory discussions, contextual inquiries, and observation, according to the specific needs within the specialised medical domain. Our contribution to the user research topic of appropriate tailoring and selection of user research methods is based on long-term case studies and hands-on experience. Based on domain experts’ needs and limitations, especially in the medical education and research field, research methods must be tailored to their specific needs to ensure the methods yield meaningful and valid results. This paper provides an overview of relevant user research methods and their adaptation in the medical field, which can serve as an empirical guide for future UCD applications

in this domain.

From Data to Dialogue: Rethinking Feedback Delivery Styles of Digital Twin in Healthcare Contexts. Yinchu Li and Regina Bernhaupt (Eindhoven University of Technology, Eindhoven, Netherlands; RMIT University, Melbourne, Australia; ruwido Austria GmbH, Neumarkt, Austria)

Human Digital Twins (HDTs) are seen as promising technologies to support health-related decision-making, provide real-time support, and simulate future scenarios. While much research has focused on improving modelling accuracy and predictive performance, particularly for clinical use, there is a growing interest in applying HDTs in everyday contexts to help individuals manage health risks and adopt healthier habits. However, through our research into how HDTs might support individuals in daily life, we identified a gap between technology-driven one-way feedback and user expectations.We reflect on two studies: a workshop study in which participants used physical and sensory materials to design how they want to receive predictive health information; and a conceptual study proposing a set of relational roles (e.g., coach, confidant, shadow) that HDTs might adopt, each implying distinct feedback styles and levels of user control. These explorations revealed a recurring topic: people do not want to passively receive predictive data from an HDT, but expect to control how, when, and in what form feedback is accessed and interpreted. Thus, we argue that HDTs should shift from one-way data delivery toward meaningful, contextual, and conversational interactions.

Usability Testing Procedure for Distraction Evaluation of In-Vehicle Digital Interfaces. Róbert Junas and Miroslav Laco (Slovak University of Technology, Bratislava, Slovakia)

In-vehicle infotainment systems have become more prevalent in the past few decades. The usability of the in-vehicle interfaces and the distraction patterns related to their usage while driving are serious research topics nowadays. Usability testing of in-vehicle interfaces poses a significant challenge. The usability testing procedure should involve advanced metrics to provide insight into the driver’s distraction patterns caused by various systems to prevent serious accidents. This paper proposes a novel usability testing procedure with a specific setup for a simulator-based driving environment. Our usability testing procedure and setup are proposed with a focus on the ease of integration with the existing standards in simulator-based usability testing of in-vehicle interfaces. We build our proposal upon the eye-tracking technique to accurately measure the driver’s distraction. We bring up novel metrics and discuss their interpretation and visualisation techniques to evaluate the distraction patterns of the in-vehicle interfaces as a part of our procedure proposal. Our proposed usability testing procedure underwent a feasibility study in the form of a case study on simulator-based usability testing of the in-vehicle system with 5 participants. The discussed results of the feasibility study point towards the usefulness of our proposed approach for evaluating the distraction patterns of the in-vehicle digital interfaces.

Session B: Smart and Secure Digital Futures

Enabling Cyber Security Education through Digital Twins and Generative AI. Vita Santa Barletta, Vito Bavaro, Danilo Calvano, Antonio Piccinno and Davide Pio Posa (University of Bari Aldo Moro, Bari, Italy)

Digital Twins (DTs) are gaining prominence in cybersecurity for their ability to replicate complex IT (Information Technology), OT (Operational Technology), and IoT (Internet of Things) infrastructures, allowing for real-time monitoring, threat analysis, and system simulation. This study investigates how integrating DTs with penetration testing tools and Large Language Models (LLMs) can enhance cybersecurity education and operational readiness. By simulating realistic cyber environments, this approach offers a practical, interactive framework for exploring vulnerabilities and defensive strategies. At the core of this research is the Red Team Knife (RTK), a custom penetration testing toolkit aligned with the Cyber Kill Chain model. RTK is designed to guide learners through key phases of cyber-attacks including reconnaissance, exploitation, and response—within a DT-powered ecosystem. The incorporation of Large Language Models (LLMs) further enriches the experience by providing intelligent, real-time feedback, natural language threat explanations, and adaptive learning support during training exercises. This combined DT–LLM framework is currently being piloted in academic settings to develop hands-on skills in vulnerability assessment, threat detection, and security operations. Initial findings suggest that the integration significantly improves the effectiveness and relevance of cybersecurity training, bridging the gap between theoretical knowledge and real-world application. Ultimately, the research demonstrates how DTs and LLMs together can transform cybersecurity education to meet evolving industry demands.

GroupSense: Capturing Group Mood Dynamics through Behavioral Analysis and Interaction Patterns in Physical Spaces. Tzu-Hui Wu, Sebastian Cmentowski, Jun Hu and Regina Bernhaupt (Eindhoven University of Technology, Eindhoven, Netherlands)

The physical workspace plays a crucial, yet frequently overlooked role in fostering social interactions, where digital solutions often inadequately incorporate physical contexts. Previous research has highlighted the promising potential of adaptive workspaces, emphasizing the importance of maintaining privacy and facilitating group-controlled interventions to foster group well-being. In this paper, we propose a future-oriented vision for an interactive system designed to leverage human interaction patterns and observable group behaviors to dynamically assess group states and promote effective group interaction. Specifically, we want to explore how group mood can be captured across three dimensions: (1) mood type, inferred through posture analysis; (2) mood intensity, identified by activity levels; and (3) mood uniformity, recognized through mimicry behaviors and convergence indicators. With our research, we want to contribute to the development of socially adaptive workspaces that are capable of autonomously sensing social interactions, improving collaboration and performance, and promoting overall wellbeing within shared workplace settings.

Perspectives on Privacy and Security in Collaborative AR - Insights From an Expert Interview Study. Sarah Claudia Krings, Enes Yigitbas and Stefan Sauer (Paderborn University, Paderborn, Germany)

With a range of consumer devices available, augmented reality (AR) continues to make its way into everyday life and is also increasingly used in professional and sensitive contexts. Especially collaborative use cases, such as remote meetings or maintenance, offer potential but also move issues in the area of privacy and security into focus. To gain an overview of the current situation and deeper insights into this field, we conducted interviews with a range of domain experts. The interviews cover the general situation regarding privacy and security in collaborative AR but also consider the specific situation and issues that end-users, on the one hand, and developers, on the other, face. The interviews revealed that end-users are almost completely dependent on the developers or device manufacturers regarding their options for protection. Developers, on the other hand, mostly do not put much focus on privacy and security, also due to external constraints, such as employers treating the topic as an “afterthought” and not budgeting enough time for it. The experts suggested offering better software support to developers to enable them to include privacy and security with less overhead as a solution. It was also suggested that a general rise in privacy and security awareness could improve the overall situation. 

Session C: Human-AI Interaction and Cognitive Models for Future Systems

Conceptualize Digital Twins as "Time Machine" Analogy. Yaxin Zheng, Harm van Essen, Scott Mitchell, Liam Fennessy, Laurene Vaughan and Regina Bernhaupt (Eindhoven University of Technology, Eindhoven, Netherlands; RMIT University, Melbourne, Australia)

This paper introduces a conceptual tool for designers engaged in Digital Twins (DTs) development. We develop the "Time Machine" analogy, a conceptual tool for navigating the DTs’ temporal dimensions: past (historical data), present (real-time monitoring and actions), and future (prediction and optimization). Through this tool, the following design considerations are formulated for designers when they design DT interactions in temporal interfaces: 1) Present information from all time frames (Past, Present, Future); 2) Connect data across time frames; 3) Allow scheduling and time management of events; 4) Encourage active user interaction and navigation through time frames. Viewing DTs as a "Time Machine" could allow designers to bring past data insights, present status, and future outcomes together, providing a

conceptual tool to imagine, visualize, and manipulate the temporal interface of DT systems. We also propose a set of guiding questions that considers how people relate to DTs and how this is taken into account when designing the interaction between users and DTs.

Conversational Agent Integration in Medical Applications: A Case Study. Adriana Pikartová, Martin Dubovský, Miroslav Laco and Vanda Benešová (Slovak University of Technology, Bratislava, Slovakia)

With recent popularity of large language models (LLM), a new topic for research in the area of Human-AI interaction (HAI) has emerged - conversational agents. Design and development of conversational agents present both design challenges regarding Human-Computer Interaction (HCI) and application challenges regarding technology complexity and limitations. One domain that could greatly benefit from such agents is medical imaging, where there is the potential to simplify and

automate tedious and repetitive tasks performed by medical experts. In this paper, we focus on the case study of design and integration of an conversational

agent into a medical image annotation tool. We propose an innovative design approach for integration of a conversational agent into existing user interface (UI) using user-centered approach, that includes research related to the assistant’s scope, design of user dialogues, user interface and overall interaction. According to the proposed approach, we integrated the assistant into a Primary Ciliary Dyskinesia (PCD) annotation tool and evaluated it through user testing with five participants for qualitative evaluation of the designed approach. Our work explores the potential of LLM-powered conversational agents, emphasizing the importance of conformity with the existing HAI principles. It provides insights into the design process and integration of conversational agents, tailored for but not limited to the medical domain.

Enhancing LLM-Based Information Seeking: Guiding and Retracing Search Trails with ChatInVis. Xabier Garmendia, Oscar Diaz and Marco Winckler (University of the Basque Country, Spain; Université Côte d'Azur, France )

Information seeking refers to the process of searching for information to satisfy specific needs or answer particular questions. As users navigate this journey, they leave behind search trails: the sequences of steps, decisions, and interactions that document their exploration. These trails capture how users navigate through various sources and ideas, revealing the progression of their reasoning and discovery. Preserving and organizing these search trails contributes to improved learning, supports decision-making, and provides a head start for future information retrieval. While Large Language Models (LLMs) have significantly advanced information-seeking

processes by providing rapid access to vast amounts of knowledge, their traditional text-based interfaces pose challenges. These challenges include difficulty in revisiting previous information and a lack of structural organization, which hinders users from efficiently maintaining and reviewing their search trails. To address these challenges, we propose a new interactive mind map-based interface, implemented through ChatInVis, a browser extension that expands the interface and capabilities of an existing mind-map visualization tool to support LLM-based information seeking interactions. ChatInVis incorporates analytic provenance, enabling users to visually document and retrace their reasoning during the search process. Additionally, it offers exploratory guidance to facilitate richer and more structured search activities. In a user study with 20 participants, we evaluated the ChatInVis for its usability and effectiveness in facilitating exploration and retracing search paths. The study focused on usability aspects and insights that participants could get while using the tool to explore unknown topics. The results indicate that our approach allows users to revisit their search trails, fostering reflection on discoveries and identification of connections. Additionally, it enables users to navigate complex information, deepen their understanding, and perform searches more efficiently. 

User Mental Models as a Lens for Evaluating Explainable-AI System Outputs. Carmelo Ardito, Maria Luigia Natalia De Bonis, Tommaso Di Noia, Eugenio Di Sciascio, Giuseppe Fasano and Angela Lombardi (Polytechnic University of Bari, Bari, Italy)

Traditional evaluation methods for Explainable AI (XAI) often emphasize perceived clarity or task performance, overlooking the cognitive processes by which expert users construct and refine their understanding. In this work, we propose a mental model-based evaluation framework that, grounded in cognitive science and Human-Computer Interaction (HCI), defines five analytical dimensions to assess how end users engage with XAI systems. Our goal is to support the design and evaluation of XAI interfaces that enable informed, trustworthy, and reflective decision-making in professional practice. we illustrate this perspective outlining how the mental model-based framework could be used to assess explanation effectiveness of Brain Age Predictor, an interactive interface for brain age prediction based on structural Magnetic Resonance Imaging (MRI) data.