Ana Müller

• Skilled - Sozioempathische KI-basierte Dialoge
  Kommerzielle Sprachsysteme haben in unserem Alltag an Relevanz gewonnen und verstehen einfache Fragen, führen jedoch keine tiefergehenden Dialoge mit Nutzer:innen. Von Code zu Empathie - Das Forschungsprojekt „Skilled“ verbessert den Einsatz sozialer Roboter und digitaler Avatare im öffentlichen Raum durch die interdisziplinäre Erforschung sozioempathischer Mensch-Maschine- Interaktion. Die im Rahmen der integrativen und agilen Forschung und Entwicklung gewonnenen Erkenntnisse werden in den Innovationslaboren des Cologne Cobots Labs, Reallaboren und in Langzeit-Feldsettings an Verkehrsknotenpunkten, in Museen und der Stadtverwaltung erprobt. Hier wird KI nicht nur klüger, sondern auch menschlicher, denn das entwickelte System begegnet Nutzer:innen stets auf Augenhöhe.
  Skilled

31st IEEE International Conference on Robot and Human Interactive Communication, RO-MAN 2022; Napoli

• Investigating Gender-Stereotyped Interactions with Virtual Agents in Public Spaces
  Müller, Ana; Penkert, Lydia; Schneider, Sebastian; Richert, Anja, 2022, Hg.: Institute of Electrical and Electronics Engineers Inc.
  Current research on the impact of gender appearance in virtual agents and social robots highlights the danger of transmitting and solidifying existing gender stereotypes. To investigate gender-stereotyped interaction at public spaces in dependency of virtual agents gender, we varied the gender of a virtual agent at a metro station. We used an ethnographic study approach, combining a two-day behavior observation with semi-structured interviews with descriptive and qualitative system log analysis of four weeks. Our results show that topics of conversation differ in dependency of the virtual agents gender: the male virtual agent was asked about topics such as brothels, drugs and alcohol and insulted frequently, while the female one was asked for relationship status or about flirting.
  IEEE Explore

32nd IEEE International Conference on Robot and Human Interactive Communica-tion (RO-MAN)

• I’ll Get By with a Little Help from AI – Exploring New Perspectives for Non-Engineer Scholars with ChatGPT
  Müller, Ana; Richert, Anja, in press, Hg.: In 2023 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)
  Large-scale language models (LLMs), like ChatGPT, have revolutionized text generation, prompting concerns about evaluation, ethics, and interdisciplinary impact. This paper advocates for interdisciplinary collaboration by empowering non-engineering scholars to shape technology without coding expertise. Using a practical example from social robotics research, it demonstrates the potential of LLMs in the first steps of programming a Furhat robot for an empirical study. It highlights the opportunities of Human-Artificial Intelligence (AI) collaboration, emphasizing different ways of problem-solving through dialogue between humans and AI. The paper aims to alleviate the fear of missing out for non-engineers, fostering a sense of inclusion.

32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN)

• No One is an Island—Investigating the Need for Social Robots (and Researchers) to Handle Multi-Party Interactions in Public Spaces
  Müller, Ana; Richert, Anja, November 2023, Hg.: 32nd IEEE International Conference on Robot and Human Interactive Communication (RO-MAN). IEEE RO-MAN, Busan.
  Social robots are increasingly used in public spaces, but they still struggle to handle complex social dynamics and interactions with multiple users. This study assessed the interaction between visitors and a Furhat robot connected to an artificial intelligence (AI)-based dialogue system in an oceanographic museum. Our findings from a video analysis of 176 interactions highlight the importance of understanding social dynamics in social robotics research. At the same time, it suggests that social robots are still limited in their ability to handle multi-party interactions (MPI) and understand complex social dynamics. Those limitations are due, in part, to technical constraints, as well as the need for further research on the social and cultural factors that shape human-robot interactions.
  IEEE Explore

EVA Berlin 2023 Conference

• May A(I) Help You? Assessing User Experience from a Field Study on Human-Robot Interaction in a Museum
  Schiffmann, Michael; Müller, Ana; Neumeister, Anke; Richert, Anja, in Veröffentlichung, Hg.: EVA Berlin 2023 Conference
  : Social robots aim to provide assistance across diverse applications through seamless natural language communication. The effectiveness of this communication largely hinges on the sophistication of the underlying dialogue system technology within the robot. To assess the practicality and reception of a social robot serving as a museum guide, we deployed a Furhat robot at the Ocean Museum Germany, OZEANEUM, Stralsund. Based on a partly standardized questionnaire with N = 141 participants, we analyzed user expectations and the perceived performance of social robots in the public service contexts. The study shows that users expect a proactive conversational design that leads to informative dialogs that are backed up by extensive knowledge about the museum and museum topics in the broadest sense. Concerning turn-taking, it becomes clear that the question-answering interaction strategy, which is typical of many commercial robots, does not match the expectations of fluid and natural communication that are set by the anthropomorphic design.

Tagungsband

• Exploring Beyond the Exhibits – Creating Knowledge for Social Ro-bots in Museums
  Müller, Ana; Schiffmann, Michael; Neumeister, Anke; Richert, Anja, 27.12.2023, Hg.: S. Thiel und J. Bernhardt, Hrsg., transcript Verlag
  Technological advancements are integrating social robots into public spaces. The scientific community has consequently become aware that relying on laboratory studies may not be the best approach to developing (social) robots, particularly when they are meant to interact with humans. Studies of human- robot interaction in real-world settings, such as museums, are deemed more beneficial in under- standing the complexities of these interactions (Sabanovic et al. 2006). Our goal in conducting this study was thus to better understand the requirements with respect to social robots in public spaces and to make them as representative of real-world conditions as possible by drawing on actual users, systems, and environments for real-world tasks. Our analysis of user utterances offers insights into identi- fying user expectations as well as system limitations and weaknesses that can be addressed through further development and training.
  https://www.transcript-publishing.com/978-3-8376-6710-3/ai-in-museums/?number=978-3-8394-6710-7

• Empathische, multilinguale KI-Systeme
  Manche Revolutionen kündigen sich lange im Voraus an, während andere scheinbar über Nacht geschehen. Die rasanten Fortschritte im Bereich generativer Künstlicher Intelligenz (KI) der letzten Monate lassen vermuten, dass auch Systeme wie soziale Roboter und Avatare in naher Zukunft den Alltag von Menschen noch vielfältiger unterstützen können. Nichtsdestotrotz zeichnet sich parallel zu dem Wettlauf in der Entwicklung neuerer, feinerer und schnellerer KI-Systeme ein immer größer werdender Forschungsbedarf ab. Es gilt, generative KI-Systeme tiefergehend und interdisziplinär zu erforschen, um mit ihrer Hilfe soziale Innovationen zu realisieren. In dieser Session erhalten Sie einen Einblick in den aktuellen Stand der Forschung und Technik im Feld generativer KI-Modelle und ihr Potenzial, dank crosslingualer Fähigkeiten soziale Innovationen zu gestalten. Wir laden Sie ein, mit uns über Chancen und Risiken der Modelle zu diskutieren und freuen uns auf spannende Perspektiven.
  08.11.2023
  BMBF Forschungstour "Ich, Zukunft und digitale Sprachverarbeitung"
• No One is an Island - Investigating the Need for Social Robots (and Researchers) to Handle Multi-Party Interactions in Public Spaces
  Presented at the 32nd IEEE International Conference on Robot and Human Interactive Communication (IEEE RO-MAN 2023)
  30.08.2023
  IEEE RO-MAN 2023
• I'll Get By with a Little Help from AI – Initial Exploration of New Perspectives for Non-Engineer Scholars with ChatGPT
  Poster presented at the 32nd IEEE International Conference on Robot and Human Interactive Communication (IEEE RO-MAN 2023)
  30.08.2023
  IEEE RO-MAN 2023
• No One is an Island – Investigating the Need for Social Robots (and Researchers) to handle Group Interactions in Public Spaces
  Doktorand*innen-Kolloquium des Promotionskollegs NRW - Abteilung Medien und Interaktion
  03.03.2023
  PK NRW
• Investigating Gender-Sterotyped Interactions With Virtual Agents in Public Spaces
  Presented at the 31st IEEE International Conference on Robot and Human Interactive Communication (IEEE RO-MAN 2023), Neapel Italy
  01.09.2022
  IEEE Explore
• Potenziale des Affective Computing bei der Erzeugung digitaler Ver-haltensdaten am Beispiel der Erfassung von Vertrauen.
  gehalten auf der Frühjahrstagung der Sektion Methoden der empiri-schen Sozialforschung der Deutschen Gesellschaft für Soziologie: Po-tenziale und Herausforderungen von digitalen Verhaltensdaten in der empirischen Sozialforschung, digital
  24.03.2022
  GESIS
• Evaluation KI-basierter sozialer Roboter und Avatare in einem Museum
  gehalten auf dem Workshop zu methodenpluralen Forschungsdesigns der Mixed-Method Sektion der Deutschen Gesellschaft für Soziologie, digital
  12.11.2021
• Skilled: Interdisciplinary, mixed methods investigations towards Socioempathic AI-based Dialogues.
  Posterpräsentation bei der 12th Media Psychology Conference, Aachen
  09.09.2021
  MediaPsych

• Promotionskolleg NRW
  Promovendin in der Abteilung "Medien und Interaktion"
  
  PK NRW