Abstract:

AI-driven chatbots are now widely used in language learning to provide personalized practice, increase engagement, and support learner autonomy. However, because chatbot interactions are typically individual and relatively new to most instructional settings, there is still limited understanding of how learners manage extended conversations with these tools and how effectively they remain focused on task goals. This study addresses that gap by analyzing interaction data from engineering students who used a Poe-based chatbot designed to act as an investor evaluating their innovation proposals. The activity was embedded in a course where students prepare for live, assessed investor-style simulations and must demonstrate clarity, relevance, and professional communicative competence.

Drawing on the concept of field from systemic functional linguistics—defined as “what the conversation is about”—the study investigates field shifts, or moments when learners drift away from the central topic in ways that disrupt coherence and communicative effectiveness. Through qualitative analysis, the study identifies several recurring types of field shifts. Among the most prominent were overly general responses (“Zooming Out”), answers that broadly missed the main point of the investor’s question (“Wide Miss”), and fragmented, hard-to-follow contributions (“Blurred Shot”). These shifts frequently led to weakened argumentation, incomplete answers, and missed communicative opportunities, ultimately limiting students’ success in the simulated investor dialogue.

This session will be of particular interest to language instructors, communication specialists, and instructional designers working with AI-mediated learning environments. By examining where and why students lost focus in chatbot-driven tasks, the presenter offers insights into how field shifts can be recognized and used as diagnostic indicators of communicative difficulty. The findings highlight the pedagogical value of chatbot interaction data for supporting more focused, purposeful, and professionally aligned learner performance.

Keywords:

AI-mediated learning, Chatbots, English for Purposes of International Communication, Systemic Functional Linguistics.