Honorable Mention at MuC 2026

Excited to share that the submission, “Beyond Manipulation: How Users Perceive Harmful AI Chatbot Interactions,” by Vanessa Budde and Eran Toch, has received a Short Paper Honorable Mention at MuC 2026! 🏆

The paper is available here.

As AI chatbots become part of more areas of everyday life, concerns about harmful interactions are growing. Some harms involve manipulation, such as biased framing or attempts to steer users’ behavior. Others are more relational, involving boundary violations, inappropriate responses, or social discomfort. Our work asks whether users actually experience these as distinct forms of harm.

In a mixed-methods study with 100 participants, we examined people’s recollections of positive, inappropriate, and manipulative chatbot interactions. The results suggest that users distinguish between different dimensions of harmful interactions. Manipulative encounters were more strongly associated with experiences of being deceived, controlled, or taken advantage of, while inappropriate interactions were more often associated with relational contexts and social discomfort.

These findings can help us better understand what makes an AI interaction harmful from the user’s perspective, with implications for designing trustworthy AI systems and emerging regulatory frameworks such as the EU AI Act.