EnglishBot: An AI-Powered Conversational Interface for Second Language Learning

Title of the Paper

EnglishBot: An AI-Powered Conversational System for Second Language Learning

Bibliographic Information

• Subject Area: Human-Computer Interaction and Educational Technology
• Keywords: Second Language Learning, Educational Chatbot, Conversational Interface, Voice Interface, Human-Computer Interaction, Language Education, Artificial Intelligence, Adaptive Feedback, Engagement, Learning Efficiency

Research Background and Issues

• Identified Problems or Challenges:

  1. Lack of opportunities to practice spoken foreign languages. Traditional methods often rely on repeated listening and speaking of recorded materials, making it difficult for students to engage in interactive speaking practice with human partners.
  2. Students often avoid practicing speaking due to speech anxiety or lack of partners, resulting in slow progress in foreign language speaking skills.
  3. The commercial sector has begun to focus on language learning chatbots, but the educational efficiency of these systems has not been systematically evaluated.

• Significance: English speaking proficiency is crucial for students’ academic and professional development. Addressing barriers to speaking practice can help more students overcome language learning bottlenecks and improve their practical communication skills.

• Research Motivation and Related Work: The motivation behind this research is to explore how artificial intelligence, speech recognition, and natural language processing technologies can be leveraged to develop more effective language learning tools. Related studies have found that chatbots can enhance engagement and academic progress in other educational fields. However, compared to traditional language learning tools, their long-term effects and efficiency in promoting language learning remain unclear.

Solution

• Proposed Method or Solution: The authors developed a language learning chatbot called EnglishBot, which interacts with students in conversational dialogues and enhances language learning outcomes through an adaptive feedback system.

• Innovative Features:

  1. Integration of cutting-edge natural language processing and speech recognition technologies for foreign language speaking practice.
  2. Provision of interactive conversational practice and instant feedback to reduce speech anxiety and improve learning effectiveness.
  3. Experimental comparison with traditional “listen and repeat” learning systems to validate the potential benefits of chatbots.

• Implementation Steps and Key Technologies:

  1. Learning Material Design: Extracting authentic dialogue scenarios from TOEFL and IELTS mock exams to ensure academic applicability of the content.
  2. EnglishBot Interface Development: Designing the chatbot interface with two sections—learning material window (left) and conversational practice window (right).
  3. Speech Recognition Technology: Utilizing Google Chrome’s speech recognition API to ensure input accuracy.
  4. Adaptive Feedback System: Providing tiered feedback based on semantic similarity and response length of students’ spoken answers, optimized using a logistic regression model.
  5. Benchmark Design for Traditional Listen-and-Repeat Systems: Building a baseline system for experimental comparison.
  6. Experimental Data Collection and Storage: Recording user behavior and learning progress using a MongoDB database.

Research Findings

• Specific Outcomes:

  1. Compared to traditional systems, EnglishBot significantly improved students’ English speaking fluency and vocabulary retention.
  2. In free-use environments, EnglishBot users demonstrated higher learning engagement, with usage time being 2.1 times that of traditional systems.
  3. The adaptive feedback system and interactive interface design received high user evaluations.

• Experimental or Evaluation Results:

  • Under controlled usage conditions, both EnglishBot and traditional systems improved students’ vocabulary retention, but fluency showed slight differences.
  • Under free-use conditions, EnglishBot significantly enhanced fluency in script-based dialogues and vocabulary mastery but had limited effects on improving free conversation skills.
  • The learning experience provided by the chatbot was more engaging, with self-reported user engagement notably higher than non-interactive systems.

• Limitations and Future Directions:

  1. Short-Term Learning Duration: The study lasted only six days, making it difficult to assess the chatbot’s impact on long-term language learning.
  2. Limited Improvement in Free Conversation Skills: The chatbot’s ability to enhance free conversation skills in real-world scenarios requires further research.
  3. Future Directions:
     • Extend the experimental duration to study the long-term effects of the technology on language learning.
     • Enhance the chatbot’s language generation capabilities in free conversation scenarios.
     • Increase the precision of adaptive feedback for grammar and pronunciation to further ensure learning effectiveness.

Conclusion

This study experimentally validated the potential of chatbot-based English learning systems to improve student engagement and specific language skills, suggesting that such systems could become a more efficient tool for English speaking practice. Additionally, it provided recommendations for designing effective language learning systems and laid a foundation for future research.