Conversational Agents in Practicing and Assessing Interactional Competence: Current Research and Future Directions
TL;DR
An overview of artificial intelligence‐powered conversational agents (CAs) in language learning and assessment is provided, highlighting features that make CAs promising tools for scalable interactional competence (IC) practice and assessment.
Abstract
This conceptual paper provides an overview of artificial intelligence (AI)‐powered conversational agents (CAs) in language learning and assessment, highlighting features that make CAs promising tools for scalable interactional competence (IC) practice and assessment. First, we review the evolution of CAs, with a focus on their use in language education. Then, we examine current research trends in using CAs to promote and assess IC. For example, we show that CAs can elicit linguistic features, such as lexical diversity and syntactic complexity, comparable to human–human interaction. However, studies also point to limitations in CAs’ ability to elicit IC phenomena, such as requests, repair, and negotiation of meaning, as well as their capacity to provide effective feedback on IC. We discuss the pedagogical and assessment implications of these findings and conclude with suggestions for future research to enhance CA capabilities and explore how interacting with AI can help learners navigate human–machine communication—an emerging aspect of IC.