Artificial Intelligence in Foreign Language Teaching – Insights from an Interview with ChatGPT and Bard

Abstract

This study investigates the evolving metacognitive strategies employed by hyperpolyglots during a twelve-month language acquisition journey. Thirty participants, each proficient in at least six languages, were analysed through a mixed-methods approach. Key findings reveal significant shifts in strategy usage, notably a decline in structured goal-setting from 87% to 72%, juxtaposed with an increase in self-monitoring from 65% to 88% and reflection-based strategies from 47% to 81%. These results highlight a transition towards more adaptive learning approaches, emphasising real-time assessment and reflective practices as essential components of effective language learning. Furthermore, factors such as language similarity and cognitive load played crucial roles in shaping participants' strategy adaptations, with 76% employing goal rotation and 68% using cognitive cycling. These findings underscore the dynamic interplay between proficiency levels, cognitive management, and metacognitive awareness, contributing to a nuanced understanding of language acquisition among hyperpolyglots.