Ort: WH G 001

• 14:15 Uhr: Daphne Auer: Self-efficacy and social cohesion in the Information Age: An app-based approach to making civic engagement more accessible
  Social cohesion and consciously shared societal values are central pillars of strong democracies. However, the spaces and mechanisms that foster cohesion and engagement within society need to be developed, as the information age is changing the way people access information, communicate, and perceive themselves as active participants. We are developing the app “ZUMI – zum Mitmachen”, which aims to motivate people to engage locally in the physical world through low-threshold, online-initiated opportunities for participation.

   

• 14:40 Uhr: Mevre Tunca: PrivacyRisq
  Digitale Demokratie setzt voraus, dass Bürgerinnen und Bürger Vertrauen in den Umgang mit ihren Daten und digitalen Infrastrukturen haben. Gleichzeitig verdeutlichen zahlreiche Datenschutzverletzungen die anhaltende und teilweise zunehmende Gefährdung personenbezogener Daten. Obwohl rechtliche Rahmenwerke wie die Datenschutz-Grundverordnung existieren, fehlt bislang eine systematische und vergleichbare Aufbereitung realer Datenschutzvorfälle, die Organisationen bei der Bewertung konkreter Risiken unterstützt.

  Der Beitrag stellt das Projekt PrivacyRisq vor, eine Webanwendung zur strukturierten Analyse und Klassifikation von Datenschutzvorfällen. Die Methodik kombiniert Bedrohungsmodellierung mit LINDDUN-Threat-Trees und etablierten Risikoframeworks wie ISO/IEC 27005 und MITRE ATT&CK, um Vorfälle aus verschiedenen Datenquellen in quantifizierbare Datenschutzrisiken zu überführen.
  Zur Veranschaulichung wird im Vortrag ein historischer Fall, der Mord von Henry Symeonis vor rund 763 Jahren, genutzt. Anhand dieses Beispiels wird gezeigt, wie komplexe Ereignisse iterativ strukturiert und zu einem nachvollziehbaren Risikomodell weiterentwickelt werden können. Die Ergebnisse verdeutlichen, dass Datenschutzvorfälle häufig aus mehreren organisatorischen und technischen Faktoren entstehen und daher integrierte Ansätze zur Risikobewertung erfordern.

• 15:05 Uhr: Paulina Stefanović: How We Draw in 3D:A Longitudinal Study of Embodied Strategies in VR Sketching
  Virtual Reality (VR) sketching enables users to create 3D forms through full-body interaction, changing how spatial cognition, motor control, and creative decision-making come together. This PhD project investigates how people learn to sketch in VR across different drawing contexts, from controlled geometric tasks to creative, body-centric and spatially constrained scenarios.

  The first study uses a longitudinal four-session design to analyze how users construct geometric forms in VR. Future studies extend this work toward creative drawing, comparing body-centric organic sketching, design spaces, fixed-space constraints, and different creative strategies. Overall, the thesis examines VR sketching as an embodied interaction practice, with implications for designing tools that better support spatial reasoning, creativity, and learning.

   

• 15.30–15.45 Uhr: Pause

   

• 15:45 Uhr: Mareike Lisker: Between Promise and Practice: Decentralized Social Media and the Democratic Public Sphere 
  Soziale Medien sind zu zentralen Infrastrukturen demokratischer Öffentlichkeit geworden, doch unterliegen zugleich den Profitinteressen und der ideologischen Willkür privatwirtschaftlicher Akteure. Dezentrale Plattformen wie das Fediverse werden als Alternative diskutiert. Dort bestimmen tausende selbst organisierte Server ihre eigenen Regeln, ohne algorithmische Kuratierung durch kommerzielle Plattformen.

  Auf Basis einer Studie von über 15.000 Fediverse-Servern fragen wir, inwieweit dieses Potenzial bereits eingelöst wird und welche strukturellen Voraussetzungen (noch) erfüllt sein müssten, damit dezentrale soziale Medien tatsächlich zu einer demokratischeren digitalen Öffentlichkeit beitragen können.

• 16:10 Uhr: Camila Lombana-Diaz: The Code Doesn't Decide, You Do! Responsible AI Beyond the Automation Myth
  AI systems are often treated as neutral arbiters: autonomous decision makers operating beyond human accountability. This talk dismantles that myth. Every algorithmic output is the product of deliberate human choices: in data selection, objective functions, and deployment contexts. The code doesn't decide. We do. Drawing on the history of Human Factors Engineering and its role in transforming software development, this talk proposes a paradigm shift: ethical AI is not a philosophical checklist or a compliance hurdle, it is a design constraint. Just as Human Factors once reshaped software by centering human abilities, limitations, and behaviours, human rights must now shape the future of AI. If an ethical principle cannot be operationalized into a design constraint or a workflow step, it is not ethics — it is marketing.

  The talk introduces the Taxonomy of Sociotechnical Harms (Shelby et al., 2023) as a practical vocabulary for identifying, naming, and measuring harm across five categories: representational, allocative, quality of service, interpersonal, and societal. It argues that closing the "productization gap", (between abstract ethical principles and real world technical implementation) is not a technical problem. It is a human one.

• 16:35 Uhr: Elisabeth Steffen: Ctrl-F-Resist. Practices, Challenges, and Technical Needs of Civil Society Organizations for the Online Analysis of Anti-Democratic Movements
  Anti-democratic actors are increasingly using online platforms to spread ideology and mobilise support. Civil society organisations play a central yet largely unrecognised role in analysing and documenting these dynamics. This talk presents a qualitative study involving practitioners from Germany, which examines how these organisations conduct analyses in the digital sphere, the challenges they face, and their technical needs.
  
  The findings show that, due to a lack of suitable tools, the work of these organisations remains largely manual and is therefore very labour-intensive. At the same time, openness to increased technical support goes hand in hand with scepticism towards automated classification by AI models and an awareness of their limitations. Based on the empirical findings, a conceptual workflow and recommendations for the development of specific tools, as well as for research and policy, will be presented.