Conversational Argument Search Under Selective Exposure: Strategies for Balanced Perspective Access
* denotes equal contribution.
Proceedings of the 48th International ACM SIGIR Conference 2025
Conversational SearchSelective ExposureArgumentationInformation Retrieval
Can conversational search help us encounter views we would skip?
When exploring a contested issue, it is easy to follow arguments that already fit our beliefs. This selective exposure can also happen in a conversation with AI. We examined two ways to help people engage with a wider range of perspectives.
Change both the presentation and the conversation
One strategy used multiple agents to structure the presentation of different viewpoints. The other had agents ask questions to encourage deeper engagement. A 2-by-2 user study compared these strategies separately and together, allowing us to examine the role of each.
Comparison and reflection played different roles
The multi-agent setup supported broader comparison between perspectives. Agent-initiated questions encouraged deeper reflection on those perspectives. Together, they promoted more balanced access to arguments, suggesting that presentation and interaction can complement one another.
Make alternative perspectives easier to engage with
Showing another viewpoint is only part of the design problem; the conversation also needs to give people a reason to examine it. The findings concern access to and engagement with arguments in the study. They should not be read as evidence that the system eliminates bias or produces lasting changes in people’s beliefs.
Read the original abstract
Conversational argument search systems influence how users access diverse perspectives but are prone to selective exposure. To address this, we propose two strategies: an interface-level multi-agent framework that structures perspective presentation and an interaction-level questioning strategy that encourages deeper engagement. We evaluate these strategies through a 2 x 2 factorial user study, examining their impact on selective exposure. Results show that the multi-agent setup facilitates broader perspective comparison, while agent-initiated questioning fosters deeper reflection; together, they promote more balanced argument access. Based on these findings, we discuss conversational search systems to mitigate selective exposure by implementing multi-agent interactions and questioning mechanisms.