ChatLearn: Leveraging Non-Native Speaker Communication Challenges as Language Learning Opportunities

Paper Title

ChatLearn: Leveraging Non-Native Speaker Communication Challenges as Language Learning Opportunities

Publication Info

• Topic area: AI-mediated communication (AIMC) systems for language learning and multilingual communication.
• Keywords: AI-mediated communication, non-native speakers, language learning, contextual learning, incidental learning, second language acquisition, communication support, cognitive load, multilingual interaction, spaced repetition.

Background and Problem

• Problem / challenge: Existing AIMC tools focus on overcoming immediate communication barriers but fail to support continuous language acquisition for non-native speakers (NNSs).
• Significance: Addressing this gap can enhance both communication effectiveness and long-term language development, which are critical for multilingual collaboration and personal growth.
• Motivation and related work: Prior research highlights the potential of NNS-native speaker (NS) communication as a learning opportunity, leveraging theories like the input hypothesis, output hypothesis, and incidental learning. However, integrating learning into real-time communication risks cognitive overload and disruption, which current AIMC systems do not adequately address.

Solution

• Proposed approach: ChatLearn, an AIMC system that transforms NNSs’ communication challenges into language learning opportunities during real-time conversations.
• Novelty:
  1. Introduces learning-oriented features like Expression Explorer, Expression Extractor, and Contextual Review Cards.
  2. Balances communication and learning goals with lightweight, unobtrusive support.
  3. Empirically demonstrates improved expression recall without significant cognitive burden.
• Procedure and key techniques:
  • Expression Explorer: Allows users to select unfamiliar expressions for translation and examples.
  • Expression Extractor: Highlights and explains native-language scaffolds in user inputs.
  • Contextual Review Cards: Resurfaces previously encountered expressions in relevant contexts for spaced and incidental learning.
  • Conducted a mixed-methods study with 43 NNS-NS pairs, comparing ChatLearn to a baseline AIMC system.

Results

• Concrete findings:
  • ChatLearn users recalled significantly more expressions (M = 3.91, SD = 1.51) than baseline users (M = 2.05, SD = 1.56; p < 0.001, d = 1.211).
  • Germane cognitive load was higher for ChatLearn users (M = 4.12, SD = 0.88) but not significantly different from baseline (M = 3.47, SD = 1.21; p = 0.069).
  • No significant differences in extraneous cognitive load or self-perceived language acquisition performance.
• Advantage over baselines:
  • ChatLearn improved recall quantity without increasing extraneous cognitive load.
  • Encouraged critical engagement with language, reducing overreliance on AI.
• Experiments / evaluation:
  • Participants: 43 NNS-NS pairs (21 baseline, 22 ChatLearn).
  • Metrics: Recall quantity, cognitive load, communication experience, and system interaction logs.
  • Tools: Welch’s t-tests, regression analysis, and structural path modeling.
• Limitations and future work:
  • Limited to text-based communication; future work should explore group or spoken interactions.
  • Focused on Chinese-English communication; applicability to other language pairs needs validation.
  • Short-term recall was measured; long-term language development requires further study.

Summary

ChatLearn is an AIMC system designed to integrate language learning into real-time NNS-NS communication by leveraging challenges as learning opportunities. It significantly improved expression recall through features like contextual review and spaced repetition, without imposing substantial cognitive burden. While it slightly reduced perceived clarity and responsiveness, this may reflect a positive shift toward critical engagement with AI support. The findings highlight the potential for AIMC systems to balance communication and learning goals, offering a foundation for future research on scalable, authentic, and personalized language learning solutions.