Conversation analysis for artificial intelligence

초록

Transforming the analytic and technological landscape, the recent rise of generative AI has sparked a surge of scholarly pursuits across a diversity of methodological terrains, and conversation analysis is no exception. With this exciting new direction also comes the need to examine the extent to which CA has advanced as a methodology for examining human-AI interaction (HAI). Based on a carefully curated corpus of “CA for AI” studies, this paper offers a brief summary of conversation analytic (CA) findings on HAI to date, examines the various methodological aspects of such studies, and provides a methodological illustration of what a CA analysis of human-AI interaction (HAI) looks like. These various attempts culminate in a reflection on a set of challenges that characterize the use of CA for investigating HAI as well as ways of navigating these challenges, thereby forming a benchmark for future research in CA-HAI in applied linguistics.