Proceedings of the 7th German Human Factors Summer School 2025

7th German Human Factors Summer School Program & Abstractband October 8th - 9th, 2025 Organizers: Johannes Kraus1 Marlene Wessels1 Elisabeth Wögerbauer2 Tim Niewalda1 1Johannes Gutenberg-Universität Mainz, Abt. Human Factors und Ingenieurpsychologie 2Johannes Gutenberg-Universität Mainz, Abt. Allgemeine Experimentelle Psychologie Part I General Information 2 3 Welcome We are happy to welcome you all to the Seventh German Human Factors Summer School 2025. The German Summer School for Human Factors is the successor of the Berlin Summer School of Human Factors which was initiated and organized from 2014-2018 by the Department of Psychology and Ergonomics, TU Berlin. The Summer School is an annual postgraduate event that is supported by the Section of Engineering Psychology of the German Psychological Society (DGPs). The intention is to provide an interactive platform that promotes the transfer and communication of interdisciplinary skills relevant to Human Factors research. Successful postgraduate applicants (Ph.D., M.Sc., and candidates) have the opportunity to present their research interests and/or current projects for critical discussion. Prominent researchers are invited to teach advanced methods and communicate state-of-the-art research from their laboratories. The 7th German Summer School for Human Factors is hosted by Johannes Gutenberg University Mainz from October 8th to 9th. We are looking forward to inspiring talks and discussions. Target audience The target audience is Ph.D. students working in Human Factors, irrespective of whether they have just started or almost finished their Ph.D. The objective of the Summer School of Human Factors is to offer a space for Ph.D. students to connect and help each other with the empirical parts and handling of other issues concerning the Ph.D. Also, the summer school will be attended by invited senior researchers and guests, further facilitating the discussions. Venue The summer school will take place at Binger Straße 14–16 (Psychologisches Institut, Johannes Gutenberg University Mainz). The Venue can easily be reached by public transport or car. 4 Information for presenters Each talk will be scheduled for 30 or 60 minutes (depending on whether it is a short or long discussion format: •During the talk, contributors can give a short introduction to their current project (8-10 mins.) and afterwards the contributer has the opportunity to freely design a workshop format to address key issues with the audience, e.g. initialize and lead a discussion with the audience or group assignments. For the workshop format, notes, pens, etc. will be provided facilitating the creative process. We want to establish a convenient atmosphere for the speaker and a lively exchange of ideas later on. •A long discussion will be scheduled for 60 minutes: Introduction (8-10 mins) and workshop (40 mins). •A short discussion will be scheduled for 30 minutes: Introduction (8-10 mins) and workshop (20 mins). All posters will be presented in a poster session (60 mins). The audience will have the opportunity to ask questions and discuss the different topics. Questions? If you want to get in touch with the organizers to discuss some ideas or upcoming questions, please get in touch via email: •[email protected] 5 Program Wednesday, 08.10.2025 08:30 – 09:00 Registration and Coffee 09:00 – 09:30 Welcome 9:30 – 10:30 Meaningful Work In the Age of AI: The Role of Technology Agency in the Experience of Meaning and Responsibility at Work Mara Hofbauer 10:30 – 11:30 Source memory for AI-vs. human-generated online content Luise Metzger 11:30 – 12:30 In-Vehicle UX in Performance Vehicles: An Experienceand Need-Oriented HMI Design Approach for Enhancing Positive Emotional States Eileen Merstorf 12:30 – 13:30 Lunch Break 13:30 – 14:30 Posters and Coffee 14:30 – 15:30 Executive Functions as Multidimensional Drivers of Human Adaptation to Emerging Technologies: A Theoretical Synthesis and Revised Taxonomy Eva Gößwein 15:30 – 16:00 Automatisierung im Führerstand - Job Design als Akzeptanzfaktor für GoA3+ Gina Schnücker 16:00 – 16:30 Coffee Break 6 16:30 – 17:30 Unlocking Transparency: Open Science Practices for Data Preparation Hanna Schindler from 19:00 Dinner at “Beim Budiker“ Address: Raimundistraße 13, Mainz 7 Thursday, 09.10.2025 09:00 – 09:30 Coffee 9:30 – 9:45 What Is Science Communication? - And Why It Matters Selina Beckmann 9:45 – 10:00 Human & Machine: Science Communication in Teaching Marlene Wessels 10:00 – 12:00 Science Communication Workshop Selina Beckmann & Marlene Wessels 12:00 – 13:00 Lunch Break 13:00 – 13:30 Interaction Space for the Teleoperation of Tram Trains Alexandra Nick 13:30 – 14:00 When the Human Factor is Needed: Psychological Mechanisms in Supervisory Control and Automation in High-Risk Work Enviroments Maike Ramrath 14:00 – 14:30 Energy-optimized headlamp distribution in urban areas Alina Waldmann 14:30 – 15:00 Concluding Session 15:00 – 15:30 Coffee Break and Conclusion 8 Presentation topics overview Mara Hofbauer Ludwig-MaximiliansUniversität München Long Discussion Meaningful Work In the Age of AI: The Role of Technology Agency in the Experience of Meaning and Responsibility at Work Luise Metzger University of Mannheim Long Discussion Source memory for AI-vs. human-generated online content Eileen Merstorf Mercedes-AMG GmbH Long Discussion In-Vehicle UX in Performance Vehicles: An Experienceand NeedOriented HMI Design Approach for Enhancing Positive Emotional States Markus Amann Honda Research Institute Europe GmbH Poster Coordinating Internal and External HMIs for Cooperative DriverPedestrian Situation Resolution Tim Kreitmaier Hochschule für angewandte Wissenschaft Landshut Poster Entwicklung eines digitalen Trainingsprogramms für Menschen mit Schwerhörigkeit Maike Lindermayr Johannes GutenbergUniversität Mainz Poster Designed to Feel Human? A Within-Subjects Study on the Influence of Anthropomorphic Design Cues on the Perception of AIDriven Chatbots 9 Melika Miralem University of Bergen Poster The Illusion of Absence: Perceptual Consequences in Controlled and Driving Contexts Tim Niewalda Johannes GutenbergUniversität Mainz Poster Mechanisms of Anxiety in Human-Robot Interactions: An affective information processing approach Insa Schaffernak Technische Hochschule Augsburg Poster Artificial Intelligence in Ophthalmology: How Hybrid DecisionMaking Can Succeed Nellie Zienert Otto-FriedrichUniversität Bamberg, IndustrieanlagenBetriebsgesellschaft mbH Poster When the Unexpected Occurs: The Influence of Human-Machine Interaction on User Reactions in Automated Vehicles Elisabeth Wögerbauer Johannes GutenbergUniversität Mainz Poster How sound properties affect preferred distances in human-drone interaction Niklas Grünewald Johannes GutenbergUniversität Mainz Poster Giving Robots a Hard Time: A Field Experiment on the Role of Robots’ Social Behavior and Social Norms for Robot Bullying 16 Giving Robots a Hard Time: A Field Experiment on the Role of Robots’ Social Behavior and Social Norms for Robot Bullying Johannes Kraus, Niklas Grünewald, Charlotte Kapell, Marlene Wessels Johannes Gutenberg-Universität Mainz [email protected] Robot bullying, defined as purposeful obstructive or harmful behavior toward robots that impedes their tasks, is a widely discussed but still little researched phenomenon. Research findings under realistic conditions are somewhat contradictory (Bartneck & Keijsers, 2020; Raab et al., 2025; Yamada et al., 2023) and contributing factors have not been sufficiently investigated. In this field experiment (N = 35), we investigated how social norms of robot bullying and social role behavior of robots affect bullying of a cleaning robot. Social norm (pro bullying vs. anti bullying) and the social role behavior (cooperativepolite vs. functional-technological (impolite)) were experimentally manipulated. Bullying was operationalized by an adaptation of the hot sauce paradigm (Lieberman et al., 1999), in which participants chose the amount of soiling liquid to be poured onto the robot’s cleaning area. In addition, trust, anthropomorphism, safety and perceived intelligence were assessed. Results showed a significantly higher tendency for bullying towards the impolite robot than the polite one. The social norm of robot bullying and the interaction between independent variables did not show a significant effect. Trust, perceived intelligence and anthropomorphism, but not safety, were higher for the polite than the impolite robot. These findings suggest that robots’ social roles affect how humans perceive and treat them - e.g. show harmful behaviors - whereas social norms, in this context, may offer limited leverage. References 17 Bartneck, C., & Keijsers, M. (2020). The morality of abusing a robot. Paladyn, Journal of Behavioral Robotics,11(1), 271–283. https://doi.org/10. 1515/pjbr-2020-0017 Lieberman, J. D., Solomon, S., Greenberg, J., & McGregor, H. A. (1999). A hot new way to measure aggression: Hot sauce allocation. Aggressive Behavior: Official Journal of the International Society for Research on Aggression,25(5), 331–348. https://doi.org/10.1002/(SICI)10982337(1999)25:5%3C331::AID-AB2%3E3.0.CO;2-1 Raab, M., Miller, L., Zeng, Z., Jansen, P., Baumann, M., & Kraus, J. (2025). Assessing pedestrian behavior around autonomous cleaning robots in public spaces: Findings from a field observation. Proceedings of the IEEE International Conference on Robot & Human Interactive Communication (RO-MAN’25). https://doi.org/10.48550/arXiv.2508.13699 Yamada, S., Kanda, T., & Tomita, K. (2023). Process of escalating robot abuse in children. International Journal of Social Robotics,15(5), 835–853. https://doi.org/10.1007/s12369-023-00987-1 18 Meaningful Work In the Age of AI: The Role of Technology Agency in the Experience of Meaning and Responsibility at Work Mara Hofbauer Ludwig-Maximilians-Universität München [email protected] Technology has been a common factor in the workplace for decades, but recent developments in artificial intelligence (AI) are changing our relation to workplace technology (Singla et al., 2025). AI is more agentic than previous technologies, meaning that it can act increasingly proactive, autonomous, and self-governed (O’Neill et al., 2022). Potential risks of interacting with highly agentic technology, like AI, in the workplace are a loss of responsibility (Sadeghian et al., 2024), and a loss of meaning at work (Kobiella et al., 2024). Feeling responsible for a task outcome and experiencing meaning at work are preconditions for work motivation (Hackman & Oldham, 1976) and are associated with wellbeing, commitment, and decreased turnover intentions (Allan et al., 2019). Thus, it is necessary to understand whether increasingly agentic technologies are diminishing the experience of responsibility and meaning at work. While some studies investigated the effect of technology agency in tasks like smart heating (Christoforakos et al., 2024), vacuum cleaning (Diefenbach et al., 2024), or text production (Draxler et al., 2024), our understanding of the effects of agency on the experience of meaningful work is limited. Therefore, we are conducting a number of qualitative and quantitative studies to first understand the experience of meaning in modern workplaces, and then test how technology agency affects these meaningful experiences. This 19 project is linking theories of meaningful work (Lips-Wiersma & Wright, 2012) to self-determination theory (Ryan & Deci, 2000, 2017) and Weiner’s theory of social cognition (Weiner, 1995), to give a comprehensive understanding of the effect of technology agency on the experience of work. Our goal is to work towards a sweet spot of technology agency, which supports people in their task fulfillment without potentially limiting the experience of meaning and responsibility at work. Description of Goal of the Discussion: The goal of the discussion is to first identify work tasks that hold meaning for a person, and second, to explore how technology agency could be operationalized for these tasks in experimental studies. Especially, how technology could appear more or less agentic to a person in the automation of these tasks. The results of this discussion will inform the experimental studies I plan to conduct for my PhD. References Allan, B. A., Batz-Barbarich, C., Sterling, H. M., & Tay, L. (2019). Outcomes of meaningful work: A meta-analysis. Journal of Management Studies, 56(3), 500–528. https://doi.org/10.1111/joms.12406 Christoforakos, L., Kolb, A., & Diefenbach, S. (2024). I did it once, but can i do it again? the role of responsibility attribution and self-efficacy in technology for sustainable behavior change. Proceedings of the 13th Nordic Conference on Human-Computer Interaction, 1–15. https://doi.org/ 10.1145/3679318.3685339 Diefenbach, S., Ullrich, D., Lindermayer, T., & Isaksen, K.-L. (2024). Responsible automation: Exploring potentials and losses through automation in human–computer interaction from a psychological perspective. Information,15(8), 460. https://doi.org//10.3390/info15080460 Draxler, F., Werner, A., Lehmann, F., Hoppe, M., Schmidt, A., Buschek, D., & Welsch, R. (2024). The ai ghostwriter effect: When users do not perceive ownership of ai-generated text but self-declare as authors. ACM Transactions on Computer-Human Interaction,31(2), 1–40. https:// doi.org/10.1145/3637875 20 Hackman, J. R., & Oldham, G. R. (1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance,16(2), 250–279. https://doi.org/10.1016/0030-5073(76)90016-7 Kobiella, C., Flores López, Y. S., Waltenberger, F., Draxler, F., & Schmidt, A. (2024). "If the machine is as good as me, then what use am i?"–how the use of chatgpt changes young professionals’ perception of productivity and accomplishment. Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems, 1–16. https://doi.org/10.1145/ 3613904.3641964 Lips-Wiersma, M., & Wright, S. (2012). Measuring the meaning of meaningful work: Development and validation of the comprehensive meaningful work scale (cmws). Group & Organization Management,37(5), 655– 685. https://doi.org/10.1177/1059601112461578 O’Neill, T., McNeese, N., Barron, A., & Schelble, B. (2022). Human–autonomy teaming: A review and analysis of the empirical literature. Human Factors,64(5), 904–938. https://doi.org/10.1177/0018720820960865 Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist,55(1), 68–78. https://doi.org/10.1037/0003066x.55.1.68 Ryan, R. M., & Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness. The Guilford Press. Sadeghian, S., Uhde, A., & Hassenzahl, M. (2024). The soul of work: Evaluation of job meaningfulness and accountability in human-ai collaboration. Proceedings of the ACM on Human-Computer Interaction,8(CSCW1), 1–26. https://doi.org/10.1145/3637407 Singla, A., Sukharevsky, A., Yee, L., Chui, M., & Hall, B. (2025). The state of ai: How organizations are rewiring to capture value. Weiner, B. (1995). Judgments of responsibility: A foundation for a theory of social conduct. Guilford Press. 21 Entwicklung eines digitalen Trainingsprogramms für Menschen mit Schwerhörigkeit Tim Kreitmaier Hochschule für angewandte Wissenschaft Landshut, Fakultät Interdisziplinäre Studien [email protected] Schwerhörigkeit zählt zu den häufigsten chronischen Erkrankungen im Erwachsenenalter. Dennoch bleiben viele Betroffene unterversorgt, was zu erheblichen Einschränkungen im alltäglichen Leben führen kann. Während es im deutschsprachigen Raum bereits einige digitale Anwendungen zum Umgang mit der Schwerhörigkeit gibt, fehlen konkrete Trainingsangebote, die eigenständig und alltagsnah von Betroffenen durchgeführt werden können. Das Projekt HörTrain verfolgt das Ziel, ein digitales Trainingsprogramm zur Unterstützung für Menschen mit Schwerhörigkeit zu entwickeln. Betroffene sollen dabei Anpassungsstrategien erlernen, die in den Alltag integriert werden können und zur Steigerung der Lebensqualität beitragen sollen. Das Trainingskonzept für HörTrain wurde zusammen mit Fachpersonen aus der Audiologie, der Hörakustik und der psychosozialen Beratung erarbeitet. Dabei wurden neben den Grundlagen für das didaktische Konzept auch erste Anforderungen an die technische Entwicklung erhoben. Die Entwicklung der Anwendung erfolgt in Anlehnung an den nutzerzentrierten Gestaltungsprozess (DIN EN ISO 9241-210 2020) als mobile App. Bereits durchgeführt wurden die Entwicklung einer Informationsarchitektur und die systematische Erstellung von vier Personas, die typische Nutzergruppen im Kontext der Schwerhörigkeit abbilden. Sie unterscheiden sich unter 22 anderem hinsichtlich der Technikaffinität, des Versorgungsstatus und der Motivation für Veränderungen. Die Personas dienen aktuell als zentrale Referenzpunkte für die inhaltliche Strukturierung und gestalterische Ausrichtung der App. Sowohl Fachpersonen als auch Nutzer werden stets in den weiteren Entwicklungsprozess eingebunden, um die Gebrauchstauglichkeit und die Relevanz der App sicherzustellen. Das Projekt befindet sich somit aktuell an der Schnittstelle zwischen Konzeption und Entwicklung. Im Fokus stehen dabei Fragen der Interaktionsgestaltung, der modularen Strukturierung von Inhalten und der Konkretisierung der funktionalen Anforderungen an das Trainingsprogramm. Parallel dazu wird die anstehende Evaluation vorbereitet, die nach der Implementierungsphase durchgeführt werden soll. Ergänzend zu diesen Themen ist ein systematisches Review zu digitalen Gesundheitsapplikationen geplant. Ziel dabei ist es, bestehende Interaktionsdesigns zu analysieren und daraus Gestaltungsmuster für die Entwicklung der App abzuleiten. 23 Designed to Feel Human? A Within-Subjects Study on the Influence of Anthropomorphic Design Cues on the Perception of AI-Driven Chatbots Maike Lindermayr, Hannah Großwieser, Johannes Kraus Johannes Gutenberg-Universität Mainz [email protected] Recent advances in artificial intelligence (AI), particularly in machine learning and natural language processing (NLP), have led to the emergence of generative chatbots that increasingly imitate human communication patterns by employing anthropomorphic design cues such as names, profile pictures, diverse communicative styles, and simulated typing behavior. Although prior studies have highlighted the persuasive potential of such technologies (Costello et al., 2024; Salvi et al., 2025), empirical evidence on the effects of anthropomorphic design cues in human-chatbot interaction remains limited. The present study addresses this gap through a within-subjects online experiment with a representative German sample (N = 221), systematically examining how anthropomorphic design cues in chatbot interfaces and dialogue styles shape user perception. Across twelve chatbot interfaces, anthropomorphic cues in the dimensions of identity, verbal, and nonverbal cues (Seeger et al., 2021) were experimentally manipulated, both in isolation and in combination, and evaluated by participants with respect to perceived anthropomorphism, trust, reliability, competence, intention to use, uncanniness, and anxiety. In addition, dispositional variables - including the Tendency to Anthropomorphize Technology (ANTEN), Propensity to Trust in Automated Technology (PTT-A), AI literacy, and general attitudes toward AI - were assessed. The findings indicate that dispositional variables, particularly ANTEN, exert a stronger influence on anthropomorphic perception than experimental manipulations of interface 24 design. The latter primarily enhanced anthropomorphic perception among individuals with low ANTEN. To intentionally foster perceived human-likeness through chatbot interface design, a combined use of multiple anthropomorphic cues - particularly verbal cues in combination with other cue categories - appears most effective, provided that individual dispositional differences are taken into account. Furthermore, anthropomorphism was moderately positively associated with trust (b = .40, SE = .04) and weakly negatively associated with uncanniness (b = -.16, SE = .04). Reliability and competence emerged as the strongest predictors of trust and partially mediated the relationship between anthropomorphism and trust. Nonetheless, these findings represent only an initial step, underscoring the need for integrative theoretical frameworks and dynamic, interaction-oriented approaches to capture the multifaceted nature of human-chatbot interaction. References Costello, T. H., Pennycook, G., & Rand, D. G. (2024). Durably reducing conspiracy beliefs through dialogues with ai. Science,385(6714). https: //doi.org/10.1126/science.adq1814 Salvi, F., Ribeiro, M. H., Gallotti, R., & West, R. (2025). On the conversational persuasiveness of gpt-4. Nature Human Behaviour, 1–9. Seeger, A.-M., Pfeiffer, J., & Heinzl, A. (2021). Texting with humanlike conversational agents: Designing for anthropomorphism. Journal of the Association for Information Systems,22(4), 931–967. https://doi.org/10. 17705/1jais.00685 25 In-Vehicle UX in Performance Vehicles: An Experienceand Need-Oriented HMI Design Approach for Enhancing Positive Emotional States Eileen Merstorf1,2, Sophia Fritz1, Vanessa Woznik1, Thomas Bäumer2 1Mercedes-AMG GmbH 2Hochschule für Technik Stuttgart [email protected] Despite the increasing relevance of experienceand need-oriented approaches in user experience (UX) research, their application to the specific context of performance vehicles remains limited. For performance manufacturers, emotionalisation plays a crucial role in conveying brand identity and addressing the needs of the target group. This segment thus facilitates the empirical investigation of positive affective user responses, providing an accessible and insightful context representative of the broader automotive market. Due to their inherent abstraction, it is challenging to meaningfully integrate psychological needs into design processes without the aid of sophisticated methodological frameworks (Zeiner et al., 2016). This causes a discrepancy between theoretical understanding and practical applicability in human-machine interaction (HMI) design. The PhD project aims to address the aforementioned gap by developing practice-oriented design aspects that render implicit psychological needs tangible and applicable for HMI design in performance vehicles (Krueger & Brandenburg, 2024). To this end, a mixed-methods approach is employed and structured into three research strands. The first research strand focuses on the systematic exploration of relevant experience categories in the context of performance vehicles (Zeiner et al., 32 be automated, similar to existing train systems. A simple "alarm" or "contact teleoperator" button was considered sufficient for passenger-initiated communication. Participants emphasized the need for multiple camera views covering the track, it’s surrounding and the vehicle cabin. Predictive computing (e.g., hazard probability estimation in shared spaces, heat maps) is considered as useful to guide teleoperators’ attention. Furthermore, reliable voice communication and quick access to field teams were deemed essential, alongside options to communicate with external persons near the tram (e.g., truck drivers blocking shared lanes). Overall, the discussion highlights concrete design implications for teleoperator workplaces, underscoring their potential to enhance safety and operational efficiency in public transport while addressing workforce shortages. 33 Mechanisms of Anxiety in Human-Robot Interactions: An affective information processing approach Tim Niewalda & Johannes Kraus Johannes Gutenberg-Universität Mainz tim.niew[email protected] Emotional states are increasingly acknowledged as an important factor in human-robot interaction (HRI) and in the formation and calibration of adjacent psychological constructs, such as trust, regarded as a prerequisite for appropriate system use. However, the specific impact of situationally experienced anxiety on trust calibration during HRI remains under-explored. Building upon a recently developed questionnaire to assess Specific Robot Anxiety (SRA), this poster presents two interrelated research paths. The first focuses on the iterative refinement and validation of the SRA measure. On this basis, the second path explores anxiety’s dynamic influence on trust calibration processes, paying particular attention to how situational affect modulates the processing of systemand context-related cues used to evaluate a robot’s trustworthiness. The poster presents preliminary ideas and an open research agenda, inviting discussion on methodological approaches, experimental designs and theoretical perspectives for exploring how affective states influence trust calibration in HRI. 34 When the Human Factor is Needed: Psychological Mechanisms in Supervisory Control and Automation in High-Risk Work Environments Maike Ramrath Work and Environmental Psychology, University of Wuppertal [email protected] In high-risk work domains such as air traffic control, crisis management, and industries with hazardous processes, advancing automation is fundamentally reshaping the role of human operators. Increasingly, humans are transitioning from hands-on controllers to supervisory monitors, overseeing automated processes and intervening primarily in response to rare but critical system events such as warnings, errors, or failure alerts. These infrequent yet high-stakes interventions place significant demands on operators’ cognitive resources. While the structural impacts of automation and supervisory control have been widely examined, the psychological processes underlying human responses to critical system information remain underexplored. The present dissertation aims to address this research gap by systematically nvestigating how professionals in safety-critical environments cognitively and emotionally process system-related critical information. The project focuses on key psychological factors, including trust, user experience, and mental workload. A multi-method approach will be employed, consisting of four studies: (1) a systematic literature review synthesizing current knowledge on the psychological effects of system-critical messages, (2) a qualitative study exploring the role of trust through practitioner interviews, paired with an online validation study, (3) a quantitative experiment examining the perception and interpretation of system errors and (4) a quantitative study on trust repair 35 mechanisms following technological failures. The anticipated contribution of this dissertation is twofold: Theoretically, it will enhance our understanding of the psychological mechanisms guiding human interactions with automated systems in high-risk settings. Practically, it will generate evidence-based recommendations for the design of socio-technical systems that foster appropriate trust calibration, reduce mental workload, and support effective human intervention during critical events. These findings aim to facilitate safer and more resilient processes in increasingly automated, highrisk work environments. 36 Artificial Intelligence in Ophthalmology: How Hybrid Decision-Making Can Succeed Insa Schaffernak Technische Hochschule Augsburg insa.sc[email protected] The use of artificial intelligence (AI) as a decision-support system in medicine holds the potential to increase diagnostic accuracy and reduce the workload of medical professionals. But how should such systems be designed and integrated into clinical practice so that they are accepted by both physicians and patients? In my doctoral project, I examine how hybrid decision-making processes (i.e., situations in which physicians receive support from AI systems in making medical decisions) can succeed in ophthalmology, taking into account psychological processes and topics related to human-machine interaction. To this end, two studies were conducted. In our interview study with ophthalmologists, it became clear that openness toward medical AI systems alone is not sufficient for their long-term integration into everyday clinical practice. Rather, a variety of factors plays a role, such as characteristics of the AI system, institutional infrastructure, adopter characteristics, and patient acceptance. Moreover, what appeared to distinguish AI systems from traditional, digital tools was the ophthalmologists’ psychological appraisal of these systems. In a second, experimental mixed-methods study, we investigated the perspective of individuals affected by AI-supported medical decisions ("decision subjects"). The aim of the experimental vignette study was to capture reactions to diagnostic decisions that involved AI in different ways. The results show that the type of AI involvement influences potential patients’ trust in the 37 final medical decision. Both the type of AI support as well as its supposed integration into the physician’s decision-making workflow affected reported trust in the medical decision. Results from sentiment analysis and qualitative coding of open-ended responses further highlight the differing perceptions of various configurations of human-AI decision-making in medical contexts, and provide suggestions for integrating AI decision support into medical consultations that do not diminish, but ideally increase trust in the medical decisions. 38 Unlocking Transparency: Open Science Practices for Data Preparation Hanna Schindler Technische Universität Berlin [email protected] Open science practices - such as preregistering hypotheses and analysis strategies or publishing datasets - are increasingly applied in the field of psychophysiology. Research in this area often follows a confirmatory approach, with clearly defined (and increasingly preregistered) hypotheses and statistical analysis plans. However, the data structures involved are typically complex and require extensive preprocessing before hypothesis testing can be performed. These preprocessing steps cannot always be fully anticipated and often depend on data quality or other unforeseen characteristics of the dataset. As a result, complete preprocessing pipelines are rarely preregistered. This is concerning given that preprocessing introduces a high degree of analytical flexibility, which can drastically impact study outcomes. In my input presentation, I will illustrate this issue using the example of my first Ph.D. study, which was a confirmatory psychophysiology study with preregistered hypotheses and statistical analysis plans (https://osf.io/8q9gm), but without a preregistered preprocessing plan. The goal of the study was to identify psychophysiological markers that can reflect changes in mental workload. We used the Multi-Attribute Task Battery - a dynamic multitasking environment simulating aviation-related tasks - to systematically induce different levels of workload. Among other measures, we recorded eye-tracking data using a video-based remote system that outputs fully accessible raw gaze data. The transformation of raw gaze coordinates into gaze metrics - such as fixations or saccades - involves multiple decisions regarding filtering, artifact rejection, and data segmentation. We applied several published and peer-reviewed 39 preprocessing pipelines to the data and found that the resulting gaze metrics varied considerably depending on the pipeline used. This observation raises two key questions: (1) What strategy should guide the selection of a preprocessing pipeline, especially when multiple valid options exist? And (2) how can such rather exploratory preprocessing decisions be transparently reported in a way that aligns with open science principles - without undermining the credibility of an otherwise confirmatory study? 40 Automatisierung im Führerstand - Job Design als Akzeptanzfaktor für GoA3+ Gina Schnücker Deutsches Zentrum für Luftund Raumfahrt e.V. gina.schnueck[email protected] Um den schienengebundenen Verkehr in Zukunft sicherer, effizienter und kostengünstiger zu gestalten, werden derzeit verschiedene Konzepte für höhere Automatisierungsstufen (Grades of Automation, GoA) im Bahnbereich diskutiert. Für GoA3 ist vorgesehen, dass Triebfahrzeugführende (Tf) das Fahrzeug nicht mehr direkt vor Ort steuern, sondern die Automated Train Operation (ATO) im Störfall als sogenannte Remote Train Operator (RTO) überwachen und im Notfall eingreifen. Diese Konzepte verändern die Tätigkeiten der Tf grundlegend und werfen sowohl arbeitsals auch ingenieurspsychologische Fragestellungen auf. Aspekte wie Performanz, Müdigkeit, Situationsbewusstsein und Belastung wurden daher in der Vergangenheit bereits untersucht. Obwohl die Akzeptanz der Tf gegenüber ATO als einer der grundlegendsten Faktoren für eine erfolgreiche Migration hin zu GoA3 betrachtet wird, bleibt dieser Aspekt in der Human Factors Forschung weitgehend unbeachtet. In dieser Arbeit wird daher untersucht, inwiefern die Einführung von GoA3 Aspekte der Arbeitsqualität von Triebfahrzeugführenden berührt und ob eine potenzielle Veränderung dieser Aspekte einen Einfluss auf die Akzeptanz gegenüber GoA3 durch die Tf hat. Hierfür wurden 66 Tf in einer Online-Vignettenstudie hinsichtlich ihres Erlebens von Veränderungen ihrer Arbeitsqualität unter GoA3 und ihrer Akzeptanz gegenüber GoA3 untersucht. Aktuell werden die erhobenen Daten ausgewertet, sodass bei der HF Summer School die Ergebnisse der Untersuchung und die daraus gewonnen Erkenntnisse besprochen werden sollen. 41 Energy-optimized headlamp distribution in urban areas Alina Waldmann L-Lab, Forviva Hella alin.w[email protected] Electric vehicles are becoming increasingly common on the roads, offering advantages in terms of sustainability and environmental impact. To maximize their driving range, manufacturers are seeking ways to reduce energy consumption. One significant area of energy use is vehicle lighting, particularly headlamp systems. Headlamps primarily serve to enhance road users’ safety at night by expanding the visual field of the driver and improving the detectability of other road users and potential hazards. In urban environments, street lighting should contribute to overall visibility. However, despite findings in several research papers, headlamps and streetlights are still developed independently, without mutual consideration of their combined lighting effects. This lack of coordination can lead to diminished visibility: overlapping light sources may reduce contrast, causing objects to blend into the background and making detection more difficult. Based on these findings, previous approaches have proposed static headlamp distributions accounting for different levels of ambient light. However, these fail to adapt to the dynamic nature of urban driving conditions. This thesis aims to identify adaptive headlamp distributions that provide only as much light as necessary. The distributions dynamically change based on factors such as vehicle speed, ambient lighting, and the presence of other road users. Thereby, they should minimize energy use while maintaining safety. Since subjective safety assessments vary across individuals and lack realtime applicability in a dynamic driving condition, objective methods will be