Teaching foreign language with conversational AI: Teacher-student-AI interaction.

Language Learning & Technology Overview

  • Article Information:

    • Title: Teaching foreign language with conversational AI: Teacher-student-AI interaction.

    • Authors: Hyangeun Ji, Insook Han, Soyeon Park.

    • Source: Language Learning & Technology, June 2024, Volume 28, Issue 2, ISSN 1094-3501, Pages 91–108.

    • DOI: https://hdl.handle.net/10125/73573

  • Abstract:

    • Focus on conversational artificial intelligence (CAI) in foreign language classrooms.

    • Utilized Google Assistant to explore interactions among teachers, learners, and CAI.

    • Employed social network and content analyses of language classes; revealed significant teacher and CAI interaction.

    • Discusses implications of CAI collaboration in teaching and outlines future research avenues.

Introduction

  • Context:

    • ChatGPT developed in 2022 by OpenAI; generates human-like responses.

    • AI implications in education include cheating concerns, plagiarism, and inaccuracy.

    • Some schools banned ChatGPT, while others advocate for its incorporation as a teaching tool.

  • Historical Use of AI in Education:

    • Previous focus on rule-based intelligent tutoring systems (Blair et al., 2007).

    • Increased investments in AI lead to advanced educational applications (Zawacki-Richter et al., 2019).

    • Systematic reviews highlight a rise in publications on AI in education from 2015 onwards.

  • Definition of Conversational AI:

    • Defined as systems mimicking human conversational abilities in text/voice (Chen et al., 2020).

  • Utility of CAI in Language Learning:

    • Addresses issues like limited authentic communication opportunities and foreign language anxiety (Teimouri et al., 2019).

    • Increases speaking and writing confidence through interaction.

  • Research Gap:

    • Limited empirical studies on CAI-integrated language classroom interactions

    • Research questions:

    1. How do language learners, teachers, and CAI interact?

    2. How do teachers facilitate interactions in a CAI-integrated classroom?

Literature Review

Sociocultural Theories in Language Learning

  • Vygotsky's Theory:

    • Emphasizes social interaction's role in cognitive development (Vygotsky & Cole, 1978).

    • Learning as a social process through collaboration with knowledgeable individuals (Lantolf, 2006).

  • Interaction Theory:

    • Communicative interaction is crucial for language acquisition (Long, 1996).

    • Engaging in communication breakdown negotiations aids vocabulary and structural awareness.

  • Implications:

    • Teacher interactions shape classroom discourse and influence student participation and engagement (Thoms, 2014).

Analytic Approaches to Classroom Interaction

  • Social Network Analysis (SNA):

    • Visualizes interaction patterns and provides insights into classroom dynamics (Carolan, 2014).

    • Studied student interactions in both online and in-person settings (Zheng & Warschauer, 2015).

  • LIMITATIONS:

    • Limited understanding of CAI as a classroom actor and its collaborative influence in interactions.

Conversational AI in Language Learning

  • CAI as a Learning Aid:

    • Potential to enhance language acquisition by providing individualized feedback and speaking practice (Hsu et al., 2021).

    • Reduces teachers' workload while enhancing learning opportunities (Chaudhry & Kazim, 2021).

  • Innovative Use in Classrooms:

    • Studies explored IPAs like Google Assistant in language classrooms (Dizon, 2017; Moussalli & Cardoso, 2020).

  • Research Findings:

    • CAIs recognized non-native pronunciations, and learners developed strategies for interaction (Moussalli & Cardoso, 2020).

Methodology

  • Study Design:

    • Intrinsic case study with two 50-minute classes; one teacher, one CAI, and four students.

  • Participants:

    • One male and four female intermediate EFL students from South Korea; representative of typical college students.

  • Technology:

    • Google Assistant chosen for its user-friendliness and previous familiarity.

  • Data Collection:

    • Video recordings of classes and group interviews conducted after each session.

Data Analysis

Video Analysis

  • Transcribed and coded verbal and non-verbal classroom interactions.

  • Results: Overall, 229 interactions in the first class and 415 interactions in the second class were analyzed. Inter-rater reliability at 93.45% and 93.25%.

Social Network Analysis

  • Conducted using R software to analyze interactions via centrality measures.

  • Defined centrality types: degree, betweenness, closeness, eigenvector.

Content Analysis

  • Focused on using Wei et al.’s teacher talk move framework to code interactions.

  • Inter-rater reliability at 81.22% and 82.17%.

Results

Classroom Interactions

  • Sociogram Results: Teacher and CAI at the center, high interconnectedness with students observed.

    • Centrality measures: Teacher decreased from 0.659 to 0.541; CAI increased from 0.267 to 0.474.

    • Indicates shifting influence dynamics from teacher to CAI over time.

Interaction Themes

  • Identified five primary interaction patterns:

    1. CAI to Student

    2. Student to Teacher

    3. Student to CAI

    4. Teacher to Student

    5. Teacher to CAI

  • Key Findings: CAI perceived as an interactive partner, while teachers adopted varying instructional strategies.

Discussion

  • Key Insights:

    • Highlighted the dynamic role of CAI and teachers in fostering language learning interactions.

    • Addressed students' emotional responses to CAI; increased familiarity noted.

  • Teacher Strategies: Increased emotional support and attentiveness to CAI responses in second class.

  • Collaboration Enhancements: Evidence showed that CAIs can assume traditional teacher roles during language learning, providing pronunciation and resources, complemented by teacher facilitation.

Conclusion

Limitations

  • Study limited by short duration; requires broader analysis for generalized findings.

Future Research Directions

  • Emphasize long-term studies and in-depth interviews with diverse learners to better understand the CAI dynamics in classrooms and impacts on student achievement.

Significance of Study

  • Offers empirical insights into CAI contributions to language learning while emphasizing the necessity for human-AI collaboration rather than replacement.

References (Selected)

  • Alharbi, H. A. (2015). Improving students’ English speaking proficiency in Saudi public schools.

  • Baker, R. S. (2016). Stupid tutoring systems, intelligent humans.

  • Dizon, G. (2020). Evaluating intelligent personal assistants for L2 listening and speaking development.

  • Hsu, M. H., Chen, P. S., & Yu, C. S. (2021). Proposing a task-oriented chatbot system for EFL learners speaking practice.

  • Vygotsky, L. S., & Cole, M. (1978). Mind in society: Development of higher psychological processes.