LingoQ: Bridging the Gap between EFL Learning and Work through AI-Generated Work-Related Quizzes

Paper Title

LingoQ: Bridging the Gap between EFL Learning and Work through AI-Generated Work-Related Quizzes

Publication Info

• Topic area: AI-mediated English as a Foreign Language (EFL) learning for workplace contexts
• Keywords: EFL learning, AI-generated quizzes, LLM-based systems, work-related English, personalized learning, self-efficacy, task-based language teaching, retrieval practice, contextualized learning, microlearning

Background and Problem

• Problem / challenge: Existing EFL learning tools often rely on generic, decontextualized materials that fail to address the specific English needs of non-native speakers in professional contexts. Workers also find it burdensome to manually review and practice work-related English after hours.
• Significance: English proficiency is critical for non-native speakers in global industries, affecting their ability to perform tasks such as reading technical documents, writing emails, and understanding domain-specific terminology. Addressing this gap can improve task performance and engagement in language learning.
• Motivation and related work: Previous research highlights the benefits of task-based and situated learning, as well as context-aware content generation. However, little work has explored leveraging workers’ interactions with LLMs to create personalized EFL learning materials directly tied to their professional tasks.

Solution

• Proposed approach: LingoQ, an AI-mediated system that generates personalized English quizzes from workers’ LLM-based language queries, enabling lightweight, work-related EFL practice.
• Novelty:
  1. Design and implementation of LingoQ, which connects LLM-based language queries to automated quiz generation for contextualized EFL learning.
  2. Empirical evidence from a three-week deployment study showing sustained engagement and measurable learning benefits for beginners.
  3. Design considerations for integrating AI-mediated EFL learning into professional workflows while respecting work-life boundaries.
• Procedure and key techniques:
  1. Workers interact with LingoQuery, a desktop chatbot optimized for English-related queries (e.g., look-up, translation, proofreading).
  2. A backend pipeline generates multiple-choice questions from query-response pairs, incorporating context from screenshots and conversation history.
  3. Questions undergo iterative quality assurance for answerability and proficiency before being added to a question pool.
  4. Workers practice quizzes on LingoQuiz, a mobile app that mixes new and previously solved questions, providing immediate feedback and explanations.

Results

• Concrete findings:
  • Participants’ self-efficacy in English increased by 9.5% on average (p < 0.001).
  • Beginner-level learners showed significant gains on a TOEIC-based test, improving by 4 points on average (out of 30).
  • 4,462 validated questions were generated, with an average accuracy of 82.6% for first attempts, increasing to 92.4% upon repeated exposure.
• Advantage over baselines:
  • Rated significantly higher than prior EFL methods in sustainability (+1.0), content relevance (+1.6), and helpfulness for work tasks (+1.3) on a 5-point scale (p < 0.001 for all).
  • Expert evaluation confirmed the quality of generated questions, with F1-scores of 0.86 for answerability and 0.88 for proficiency.
• Experiments / evaluation:
  • A three-week deployment study with 28 Korean EFL workers, analyzing usage patterns, learning outcomes, and self-efficacy.
  • Expert evaluation of 30 sampled questions for quality assessment.
  • Pre- and post-study surveys and TOEIC-based tests to measure learning gains and user perceptions.
• Limitations and future work:
  • Limited to reading and writing skills; future work should include listening and speaking.
  • Focused on fill-in-the-blank questions; expanding question formats and modalities is necessary.
  • Study conducted in a Korean–English context; generalization to other languages requires further investigation.

Summary

LingoQ is an AI-mediated system that transforms workers’ LLM-based language queries into personalized English quizzes, enabling lightweight, work-relevant EFL practice. A three-week deployment study with 28 Korean EFL workers demonstrated sustained engagement, increased self-efficacy, and measurable learning gains for beginners. Expert evaluation confirmed the quality of generated questions, and participants valued the system’s contextual relevance and practicality for work tasks. Future work should expand question formats, include verbal communication skills, and explore broader language contexts. LingoQ highlights the potential of leveraging LLM interactions for personalized, task-specific language learning.