Year: 2026
Type: conference-paper
Abstract: People are increasingly turning to AI agents for advice and support, but can, as a result, be misled or develop an over-reliance upon a single AI voice. In this paper we offer a provocation that takes a playful approach to multi-agent conversations, where our embodied AI agents, each with their own unique characteristics and preferences, offer diverse opinions in a social discourse setting. Such a multi-agent approach stimulates critical thinking and moves us away from a simplistic assumption that the AI agent is always right.
Authors: Pam Briggs, Matt Wood, Dan Jackson, Kyle Montague
Institution: Northumbria University
Language: English
Subfield: Social Psychology
Field: Psychology
Domain: Social Sciences
Open Access status: hybrid
OpenAlex ID: W7224819333
