Investigating the Synonyms of Conversational Agents to Aid Cross-Disciplinary CA Research

Abstract

With the advent of artificial intelligence, the CHI community has regained a large interest in Conversational Agents (CAs). Designing CAs involves cross-disciplinary efforts, such as that of computer science (e.g., language models), psychology (e.g., cognition and emotions), linguistics (e.g., conversation design), or communication (e.g., trust, communication theory). However, CAs are named differently depending on the discipline and purpose (e.g., chatbot, embodied avatar, virtual butler). This divergence of vocabulary on CA brings challenges to researchers and designers in researching the full landscape of CA literature and sharing cross-disciplinary knowledge. We performed bibliometric and qualitative analyses to systematically assess divergent terms used for CA nomenclatures. We present 54 CA-terms and how these terms are used differently depending on the discipline and design characteristics (e.g., voice vs. non-verbal vs. text-based). Our work contributes to helping CA researchers effectively review the CA literature and build on each other's work for novel cross-disciplinary CA research.