Assessing How Users Display Self-Disclosure and Authenticity in Conversation with Human-Like Agents: A Case Study of Luda Lee

Abstract

There is an ongoing discussion on what makeshumans more engaged when interacting withconversational agents. However, in the area oflanguage processing, there has been a paucityof studies on how people react to agents andshare interactions with others. We attack thisissue by investigating the user dialogues withhuman-like agents posted online and aim toanalyze the dialogue patterns. We construct ataxonomy to discern the users' self-disclosurein the dialogue and the communication authenticity displayed in the user posting. We annotatethe in-the-wild data, examine the reliability ofthe proposed scheme, and discuss how the categorization can be utilized for future researchand industrial development.