Revealing the Relationship between Discourse Safety and Support in Conversational Agents with Educational Big Data

• In Event: DELT- Revealing the Relationship between Discourse Safety and Support in Conversational Agents with Educational Big Data

Short Description

Conversational agents (CAs) have attracted increasing attention in the educational community, given their ability to generate supportive responses to students automatically. However, whether a supportive response is safe can be subject to individual, contextual, and cultural differences. This study aimed to investigate the relationship between CA safety and socio-emotional support. The results showed that supportive texts of CAs can be unsafe, yielding potential threats to creating an engaging and constructive learning environment.

Authors

• Presenter: Wanli Xing, University of Florida

• Presenter: Chenglu Li, University of Florida

• Contributor: Walter Leite, University of Florida