• News

ComAI at the ACM Conversational User Interfaces Conference 2025

The ComAI sub-project “Interfaces” is represented at the ACM Conversational User Interfaces Conference 2025 with two paper presentations and a workshop. The international conference, taking place from July 8–10 in Waterloo (Canada), is one of the leading venues for research on conversational AI, voice interfaces, and human–machine interaction.

Users as Active Co-Designers of Voice Interfaces

One of the talks is titled “Beyond the Evolution: Investigating How Users Co-Shape the Design of Conversational Interfaces” (Laura Spillner, Johanna Rockstroh, Paola Raquel Peña, Nina Wenig, Nima Zargham, and Benjamin R. Cowan). This qualitative study shows that users are not merely passive recipients but actively co-shape voice interfaces – for instance through feedback, creative repurposing, or pragmatic adaptations. The paper offers new perspectives on user-centered design in conversational AI – a central concern of the ComAI “Interfaces” project.

Shifting Boundaries in LLM-Based Design

The second contribution, “Crossing the Line? Designers’ Concerns about Ethical Boundaries in LLM-based Conversational Agents” (Nima Zargham, Vino Avanesi, Laura Spillner, and Johanna Rockstroh), addresses ethical negotiations in the design of large language model-based agents. Interviews with designers reveal blurred and at times conflicting boundaries between organizational demands, creative freedom, and personal values – particularly around anthropomorphic design, emotional simulation, and role modeling. The paper calls for deeper ethical reflection and the development of new design guidelines for LLM-based interfaces.

Workshop on Ethical LLM Personas

The conference contribution is rounded off by the workshop “Personas Evolved: Designing Ethical LLM-Based Conversational Agent Personalities” (Smit Desai, Mateusz Dubiel, Nima Zargham, Thomas Mildner, Laura Spillner). Through interactive sessions, participants prototype LLM-based agents and discuss how to design them in ethically responsible, inclusive, and socially aware ways. The aim of the workshop is to develop cross-disciplinary principles for transparent and trustworthy conversational personas.