CHIME: An Interruptive Chatbot for Supporting Marginalized Voices in Group Chats
Online group chats are often dominated by voices from particular demographic groups, making it difficult for others to contribute. This article investigates whether AI mediation can redistribute speaking opportunities through strategic conversational interventions. We present Chime, a chatbot that posts brief interventions in response to sustained conversational dominance, safety risks, and prolonged silence. In a controlled experiment with 102 participants discussing football, a male-dominated topic, Chime reduced conversational monopolization without reducing overall participation. Marginalized female participants also reported greater opportunities to speak. Further analyses suggest that Chime achieved these effects by creating conversational openings without suppressing substantive discussion. Qualitative analyses identify a Dual-Buffer Mechanism, whereby Chime’s non-human identity made interruptions more face-saving for dominant speakers while its interventions simultaneously created structural openings for marginalized speakers to enter the conversation. These findings demonstrate how AI-mediated interventions can be designed to redistribute conversational opportunities and amplify voices marginalized in social interaction.
Authors
- Bo Li (ORCID: https://orcid.org/0000-0003-4905-2744)
- L.C. Ray
- Fiona Fui‐Hoon Nah (ORCID: https://orcid.org/0000-0002-5505-7843)
- Keya Chen
- Ruochen Mao
- Jiale Li
Institutions
- Hong Kong Polytechnic University (HK)
- The University of Melbourne (AU)
- National University of Singapore (SG)
- City University of Hong Kong (HK)
- Singapore Management University (SG)
Publication Details
- Journal
- International Journal of Human-Computer Interaction
- Published
- 2026-11-17
- DOI
- https://doi.org/10.1080/10447318.2026.2759213
- Primary Topic
- Ethics and Social Impacts of AI
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- City University of Hong Kong