Enhancing Sustainable AI-Driven Language Learning: Location-Based Vocabulary Training for Learners of Japanese

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study proposes a location-based AI-driven e-learning system that dynamically generates language learning materials tailored to real-world contexts by integrating location awareness technology with AI.

Current language learning systems still face three major challenges: lack of immersion, low efficiency in fragmented learning, and heavy reliance on predefined materials. These issues prevent learners from fully engaging with the language environment, thereby affecting long-term learning outcomes, and therefore the problem solved in the manuscript is relevant.

There are suggestions and comments on the study.

 

1. Only 10 survey participants is a very small sample to justify any generalisations about the applicability or effectiveness of the approach for a wider audience. The results may not be representative and may be influenced by individual participant characteristics. A larger number of participants is needed to ensure the statistical significance of the results and to identify potential differences in content perception based on age, gender and especially language proficiency.
2. It would be useful to supplement subjective assessments with objective indicators such as the results of mastery tests, time spent studying the content, and analysis of errors made by participants.
3. The text of the article does not provide a detailed analysis of the participants' responses for each of the four key parameters (Content Relevance, Content Accuracy, Learning Motivation, Learning Efficiency). It would be useful to see the distribution of scores for each parameter, as well as participants' comments, in order to get a deeper understanding of the reasons for high or low scores.
4. There is not enough information about how to work with DeepSeek-V3. How is content generated? What prompts are used? How much control is there over the content?
5. Read the text carefully, there are minor errors:

Line 254: probably Figure 3 instead of 2.

Reference to Figure 4 in the text after the figure itself.

No reference to Table 1 in the text. It should probably be on line 110 instead of Figure 1.

 

Overall, the study presents an interesting concept and shows promising results. Further research and them extension are required to provide more reliable and objective data on the efficacy and practical application of the approach.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In overall, I enjoyed reading this article: it discusses a number of important innovations contributing to the improvements in iCALL and iMALL systems. Though the concept of location-based awareness does not seem to be novel within the context of language learning systems, the described approach incorporates a few interesting ideas  and features for presenting them to a wide audience of journal readers. I specifically like en attempt to go towards a formalized approach to assess the output of AI-based text generators.

Though not completely ready for publication yet, the article touches the important aspects of contemporary agenda of language learning and will surely attract attention of the Sustainability readers. 

Here are my remarks and suggestions that, as I hope, might be helpful to improve the work:

1. Though I acknowledge that many ideas discussed in this paper are L2 agnostic, I’d suggest to narrow the title so that it explicitly reflects the current achievements: tools for vocabulary learning + tools targeting the learners of Japanese. 
2. The authors constants MALL against the traditional classroom instruction, which looks like methodologically incorrect. MALL is a method. Class room instruction (traditional or not) is not a method but an organizational process (that actually can involve different methods and approaches including MALL, by the way).  Therefore, MALL must be contrasted ageists other methods rather than against classroom teaching.
3. Comparing the classroom scenarios agains location-based scenarios forms a right approach. However, more details are required to convince that the modelled exercises in a classroom aren’t enough to cover topical vocabulary learning. Different innovative approaches to enhance vocabulary acquisition such as bilingual resources (such as manga, specifically for learners of Japanese!), contextualized videos, movies, songs, etc., must be included to such a contrastive analysis.
4. Location awareness and immersion are examples of possible ways to contextualize and target the language learning activities. I’d recommend to position the possible location-based immersive scenarios under the umbrella of broader available studies on conceptualization and targeting in CALL, CAPT, MALL, etc.
5. It is a bit confusing, that by mid-paper, it is unclear that the discussed tools particularly address the learners of Japanese. 
6. Virtualized environments (that, again, can be efficiently used in “traditional” classroom, but not only!) might happen to be more productive from the perspective of implementing immersive scenarios modeling real-world situations. Indeed, the learner can interact with such an environment in a much more convenient time slot rather than few minutes of waiting for a train or for a dish in a restaurant. Such opportunities must not be ignored.
7. The evaluation is currently based solely on subjective experience of the users answering a quite simple questionnaire. With respect to the different proficiency levels, it would be nice to know, whether the system produces the output suiting well that different levels. Also, are there any objective metrics that can be used to assess the suggested approach for vocabulary learning?
8. Unfortunately, the paper does not have any illustrations of the working prototype (screens, interaction logs, etc).
9. For the future work section, I’d suggest to take a closer look to the approaches to language feature visualization that can be helpful to improve the system interface and to produce more instructive MALL feedback to the learners.

Minor issues:

1. Fig. 2 seems to be too superficial for a “system architecture”, for me.
2. Fig. 3, please check the spelling of “restaurant”.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Learning from Daily Context: Location and AI-Driven Immersive Language E-Learning

1. Explicit Statement of the Research Gap: Include a direct statement that clearly outlines the research gap.
2. Figures and Tables Referencing and Description. Enhance the descriptions and discussions surrounding each figure and table to ensure they add value and clarity for the readers. Specifically, Figure 3 and Table 7 are mentioned in the text, but the descriptions or discussions surrounding them are not detailed. It would be beneficial for readers if you could provide a more comprehensive explanation of what these figures and tables depict and how they relate to the discussion or findings presented.
3. User Survey (Section 5.3): The user survey is well-structured and covers essential aspects such as content relevance, accuracy, learning motivation, and efficiency. However, the description of Figure 6 seems inconsistent with its caption, which mentions "Overall system architecture." Please ensure that the figure caption accurately reflects the content it represents, which should be the survey results rather than the system architecture.
4. Benefits of Dynamic Content Generation: The methodology section outlines the benefits of dynamically generating adaptive content. It discusses how this approach reduces reliance on static resources, improves accessibility, and supports sustainable learning practices. These points are evaluative and interpretative, making them more suited to a discussion section. To improve the clarity and flow of the manuscript, I recommend separating these discussion elements from the methodology section. Consider creating a distinct discussion section following the evaluation of the results. This new section could focus on interpreting the effectiveness of your methods, their impact on learning outcomes, and their potential for fostering lifelong learning. This structural change would help readers clearly differentiate between how the study was conducted (Methodology) and the implications of these methods (Discussion).
5. Insufficient Detail on Prompt Engineering: While the manuscript mentions prompt engineering, the explanation is somewhat superficial. The prompts are described in tables, but there is no discussion of how these prompts were optimized or tested for different scenarios.
6. Limited Exploration of Cultural Context: The system generates content based on location, but there is no discussion of how cultural context is incorporated. Language learning is deeply tied to culture. Understanding and integrating cultural nuances directly enhance the relevance and effectiveness of language learning.
7. Broader Implications for Sustainable Education: The paper touches on sustainable education, but this could be expanded. Discussing how the proposed system aligns with broader sustainability goals in education, such as reducing resource consumption or promoting equitable access, would add depth to the discussion.

Comments on the Quality of English Language

Some grammar lapses are observed. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

Here there are some specific improvements should the authors consider regarding the methodology and further controls should be considered

More detailed participant selection criteria and sample analysis:

• The study does not clarify how participants were selected (e.g., selection criteria, language proficiency levels, and educational background).
• More details on age groups, prior language learning experience, and the number of participants should be included to ensure the generalizability of the findings.
• A larger and more diverse sample is recommended to validate the approach across different learner demographics.

Incorporate more detailed statistical analysis:

• While quantitative results are provided, it is unclear whether statistical tests such as ANOVA or T-tests were conducted to determine the statistical significance of differences.
• Including p-values and confidence intervals will strengthen the credibility of the results.

Consider environmental factors influencing learning:

• The study does not address how external variables, such as background noise, time of day, or social interactions, might impact learning effectiveness.
• The methodology should examine how different learning environments (e.g., indoor vs. outdoor) affect user engagement and comprehension.

More in-depth AI model evaluation:

• The comparison of AI models (e.g., DeepSeek-V3, ChatGPT-4o) lacks a clear explanation of the criteria used for selecting the best model.
• A qualitative and quantitative assessment of each model’s accuracy, contextual relevance, and user interaction quality should be included.

Additional controls to evaluate the impact of geolocation on learning outcomes:

• It is unclear whether the study controlled for variations in location types, which could affect conclusions.
• A comparison of different location categories (e.g., libraries, train stations, cafés) should be included to determine which settings are most effective for language learning.

Examine the impact of repeated exposure on long-term retention:

• The study does not investigate how the frequency of content exposure affects knowledge retention over time.
• Follow-up assessments days or weeks later could help measure the durability of learned vocabulary and phrases.

Compare learners of different proficiency levels:

• The study should compare beginner vs. advanced learners to determine if the system benefits all levels equally.
• This could improve adaptive content personalization based on proficiency level.

Additional Controls to Consider

• Assess the impact of time spent on learning: Is there a correlation between the duration of location-based learning and language acquisition effectiveness?
• Measure user engagement and interaction levels: Do learners actively engage with the material, or is their interaction passive?
• Compare system performance with alternative learning methods: How does location-based learning compare to traditional mobile apps that do not adapt to real-world contexts?
• Evaluate geolocation accuracy and its effect on content quality: How do GPS errors influence the relevance and accuracy of generated learning content?
• Analyze cultural and social context influences: How does a learner’s cultural background impact their response to location-based learning?

By implementing these methodological refinements and controls, the study will become more rigorous, enhancing its scientific and practical impact.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I appreciate the significant efforts the authors made; my comments to the current version are as follows:

1. The title in the authors' response differs from the actual title of the manuscript, the latter being "Enhancing Sustainable Language Learning: AI-Driven, Location-Based Vocabulary Acquisition for Japanese Learners". To avoid confusions, what about "Enhancing Sustainable AI-Driven Language Learning: Location-Based Vocabulary Training for the Learners of Japanese"?
2. With respect to my comment No. 5 to the original submission, I'd suggest to discuss a bit more, what are specific features and particularities of the Japanese language and vocabulary acquisition issues that can be addressed using the approach described in the article.

Author Response

Comment 1: The title in the authors' response differs from the actual title of the manuscript, the latter being "Enhancing Sustainable Language Learning: AI-Driven, Location-Based Vocabulary Acquisition for Japanese Learners". To avoid confusions, what about "Enhancing Sustainable AI-Driven Language Learning: Location-Based Vocabulary Training for the Learners of Japanese"?

Response 1: Thank you for your suggestion. We have revised the title accordingly to ensure consistency and clarity.

Comment 2: With respect to my comment No. 5 to the original submission, I'd suggest to discuss a bit more, what are specific features and particularities of the Japanese language and vocabulary acquisition issues that can be addressed using the approach described in the article.

Response 2: Thank you for your valuable suggestion. We have expanded the discussion on the specific features of the Japanese language and the challenges of vocabulary acquisition in both the Introduction and Discussion sections (line 43~51, 597~603). In particular, we have highlighted how our location-based approach enables learners to practice Japanese honorific expressions in different real-world contexts. We appreciate your insightful feedback.

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript, given its revision, has improved a lot and satisfied my comments. 

Author Response

Thank you for your positive feedback.

We truly appreciate your insightful comments, which have greatly contributed to improving the quality of our manuscript.

Reviewer 4 Report

Comments and Suggestions for Authors

I would like to sincerely thank authors for their hard work and dedication in addressing the reviewers' comments and improving the manuscript. Your thoughtful revisions and clarifications have significantly enhanced the methodological rigor and overall quality of the study.

I appreciate the detailed responses you provided to each comment, demonstrating your commitment to strengthening the research. The incorporation of statistical validation, improvements in AI model evaluation, and your plans for future work further establish the importance and relevance of your study.

Based on the revisions made, I find the manuscript suitable for publication in its current form. Congratulations on your efforts, and I look forward to seeing the impact of your research in the academic community.

Author Response

Thank you very much for your kind and encouraging feedback.

We truly appreciate your thoughtful review and insightful comments, which have greatly contributed to enhancing the quality and rigor of our manuscript.