Generative AI as a Cognitive Co-Pilot in English Language Learning in Higher Education
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
- The term “cognitive co-pilot” is used without a clear definition, making it difficult to understand how AI is positioned in learning beyond a general assistive role. A precise definition is needed.
- The literature review lacks depth in pedagogical frameworks. It references studies on AI adoption but does not critically engage with theories on cognitive engagement, autonomy, or learning strategies in AI-assisted language learning.
- The study reports gender differences in AI adoption but does not conduct any statistical tests (e.g., t-tests or chi-square) to determine significance, making the interpretation weak.
- The sampling method is unclear and lacks justification. Differences in institutional policies on AI use are not considered, which could impact students’ perceptions.
- The factor analysis and regression results show weak loadings for some predictors (e.g., effort expectancy, social influence), yet they are retained without discussion. The study does not explain their practical significance.
- The qualitative analysis is weak, providing only surface-level descriptions without deep interpretation. There is no triangulation between qualitative and quantitative findings, making the mixed-methods approach inconsistent.
- The discussion does not critically examine AI overuse or dependency risks. It presents AI adoption positively but does not address whether students develop critical evaluation skills or blindly accept AI-generated outputs.
- The implications lack clear recommendations. The study suggests AI is useful but does not address whether universities should regulate AI use or if students need training on ethical AI engagement.
- The limitations section is insufficient. It mentions reliance on self-reported data but does not discuss response bias, potential overestimation of AI effectiveness, or institutional factors affecting adoption.
Comments on the Quality of English Language
needs improvement.
Author Response
Responses for Reviewer 1
Comment 1: The term “cognitive co-pilot” is used without a clear definition, making it difficult to understand how AI is positioned in learning beyond a general assistive role. A precise definition is needed.
Response 1: We have added the definition in the introduction section, as seen below:
In this study, we define “cognitive co-pilot” as a generative AI tool that supports and enhances learners’ cognitive processes—such as comprehension, idea generation, revision, and self-regulation—while preserving students' active engagement in meaning-making and decision-making. Rather than replacing human thinking, a cognitive co-pilot facilitates reflective interaction and scaffolds academic performance, similar to a collaborative learning partner. (Page 2, Lines 67-73)
Comment 2: The literature review lacks depth in pedagogical frameworks. It references studies on AI adoption but does not critically engage with theories on cognitive engagement, autonomy, or learning strategies in AI-assisted language learning.
Response 2: We have added two new paragraphs discussing pedagogical frameworks in the literature review section, as seen below:
Furthermore, recent studies on generative AI and learning have been mostly concerned with technology uptake and user experience but also need to be grounded in pedagogical frameworks that simulate learners' cognitive and emotional interaction with AI within learning environments. Grounded in Self-Determination Theory (Deci & Ryan, 2008), generative AI programs such as ChatGPT can enhance autonomy in self-directed learning, competence in individualized scaffolding and feedback, and relatedness in virtual conversational interaction (Du & Alm, 2024; Chiu, 2024). Yet, more recent research gives a less dichotomous account: whereas some students perceive AI to increase social presence and connectedness, others worry AI reduces human interaction—most notably in situations where collaborative learning or affective involvement are promoted (Du & Alm, 2024; Xie et al., 2024). These cross-pressures necessitate consideration of how AI technologies enable and potentially limit major motivational aspects of language learning.
Apart from motivation, Cognitive Engagement Theory (Greene & Miller, 1996) and ICAP (Interactive, Constructive, Active, and Passive) theory (Chi & Wylie, 2014) play a major role in comprehension of how students use AI to involve themselves in more superior learning processes. The ICAP model recognizes superior levels of engagement—especially constructive and interactive—are linked with better learning. Under AI-supported writing, students exercise metacognitive skills in several ways, such as planning, monitoring, and evaluation (Yao et al., 2024). Students typically read through, revise, and incorporate suggestions into their own writing, showing reflective and active processes (Chan & Hu, 2023; Yang et al., 2024). This artificial intelligence strategy aligns with self-regulated and metacognitive learning theories, which stress the agency of students to regulate their cognitive resources towards learning tasks (Lai, 2024). Overall, these educational theories highlight that generative AI applications cannot be viewed as exclusive aid technologies but instead as motivational and cognitive go-betweens that influence the way language learners learn, exert autonomy, and enjoy fruitful, self-initiated learning. (Page 4, Lines 166-191)
Comment 3: The study reports gender differences in AI adoption but does not conduct any statistical tests (e.g., t-tests or chi-square) to determine significance, making the interpretation weak.
Response 3: Thank you for your thoughtful comment. In response, we have conducted chi-square tests to assess gender differences in GenAI tool usage and integrated the results into Section 4.1, including a new paragraph and summary table, as seen below. While only Duolingo showed a statistically significant difference, this analysis enhances the rigor of our interpretation. Furthermore, the study includes exploratory factor analysis (EFA) and multiple regression analysis, which collectively strengthen the overall analytical framework and support the robustness of our findings.
To examine whether gender significantly influenced the use of different generative AI tools, chi-square tests were conducted on the five most reported applications. The analysis showed that Duolingo usage differed significantly by gender (χ² = 9.74, p = 0.002), with a higher proportion of female students (51.9%) using the app compared to male students (29.6%). However, no significant gender differences were found for ChatGPT (p = 0.090), Google Translate (p = 1.000), Grammarly (p = 0.304), or QuillBot (p = 0.805). These findings suggest that while overall adoption rates for major AI tools are largely uniform across genders, certain applications such as Duolingo may appeal differently, potentially reflecting gender-based preferences in language learning strategies. Table 3 below displays the results. (Pages 11-12, Lines 467-477)
Table 3. Chi-square results for gender differences in GenAI tool usage.
AI Tool |
χ² |
p-value |
Significance |
ChatGPT |
2.88 |
0.090 |
Not significant |
Google Translate |
0.00 |
1.000 |
Not significant |
Grammarly |
1.06 |
0.304 |
Not significant |
QuillBot |
0.06 |
0.805 |
Not significant |
Duolingo |
9.74 |
0.002 |
Significant |
Comment 4: The sampling method is unclear and lacks justification. Differences in institutional policies on AI use are not considered, which could impact students’ perceptions.
Response 4: We have added a new paragraph in the method section, as seen below:
The study employed a convenience sampling method (Etikan et al., 2016) based on the accessibility and willingness of students to participate from multiple public and private universities across Indonesia. This approach was appropriate given the study’s exploratory nature and the need to capture diverse perspectives from students with varying access to AI resources. While this method may limit generalizability, the inclusion of participants from eight institutions across different regions helped mitigate sampling bias. Additionally, we acknowledge that institutional policies on AI usage may vary, potentially influencing students' experiences and perceptions. Although this variable was not directly measured, it is recognized as an important contextual factor, and future studies should consider comparing student responses across institutions with different levels of AI integration support and guidelines. (Page 9, Lines 369-379)
Comment 5: The factor analysis and regression results show weak loadings for some predictors (e.g., effort expectancy, social influence), yet they are retained without discussion. The study does not explain their practical significance.
Response 5: We have added a new paragraph discussing these results, as seen below:
Performance expectancy clearly emerged as the strongest predictor of GenAI acceptance whereas effort expectancy and social influence demonstrated relatively weaker statistical contributions. Nonetheless, their retention in the model is supported by theoretical and practical relevance. Effort expectancy, although showing lower factor loadings, reflects usability concerns that remain critical for sustained engagement with AI tools, especially for less tech-confident users. Correspondingly, social influence, despite its modest variance explanation, captures peer and institutional norms that subtly shape technology adoption in collectivist cultures like Indonesia. Retaining these factors offers a more holistic understanding of the multifaceted dynamics influencing AI acceptance. These results emphasize that even predictors with weaker statistical weight can carry pedagogical and cultural importance, especially in emerging and context-specific technology environments. (Page 16, Lines 596-607)
Comment 6: The qualitative analysis is weak, providing only surface-level descriptions without deep interpretation. There is no triangulation between qualitative and quantitative findings, making the mixed-methods approach inconsistent.
Response 6: We provided a figure displaying the identified themes and descriptions, then we explained each theme in detail along with the sample excerpts from the participants. At the end of each explanation paragraph, we offer what we can learn from the results. We believe that these are the standards of reporting thematic analysis findings. To address your concern, we have added one new paragraph in the discussion section, as seen below:
Thematic analysis results provided richer insight into students’ interactions with AI, complementing quantitative findings on satisfaction and usage patterns. Students' reflective use of tools like Grammarly and ChatGPT for grammar refinement, content exploration, and idea generation mirrored the statistical prominence of writing assistance and research support tasks. This triangulation confirms that students are not merely passive users but demonstrate strategic engagement and critical filtering—especially when revising text or cross-checking AI translations. Moreover, students’ concerns about AI limitations and their adjustment of AI output align with the modest influence of effort expectancy and social influence in the regression results. These links illustrate the value of integrating qualitative narratives into the quantitative patterns, offering a more comprehensive and nuanced view of learner-AI interactions. The consistency between what students say (qualitative themes) and what they do (quantitative data) reinforces the credibility of findings and the explanatory strength of the mixed-methods design. (Pages 20-21, Lines 804-816)
Comment 7: The discussion does not critically examine AI overuse or dependency risks. It presents AI adoption positively but does not address whether students develop critical evaluation skills or blindly accept AI-generated outputs.
Response 7: We have added a new paragraph, as seen below:
While this research indicates the potential role of GenAI as an intellectual co-pilot for scholarly writing, language acquisition, and content searching, consideration needs to be given to the danger of over-reliance and diminished critical thinking. Although a number of students indicated revising AI-generated content actively and critically verifying translations, qualitative data also revealed instances where AI tools were being utilized as fact-based sources, especially in moments of urgency. This aligns with the balance between intellectual autonomy and productivity. Without guidance, students may be conditioned to rely on AI at the expense of their capacity for higher-level thinking, particularly in assessing bias, context relevance, or appropriateness of AI-driven content. These factors align with Malik et al. (2025) and Watermeyer et al. (2023), who cautioned that unchecked application of AI potentially destabilizes core academic skills and moral judgment. Hence, GenAI integration should be accompanied by direct teaching of critical thinking approaches and ethical application in a bid to sustain students' active engagement as reflective learners as opposed to acting as passive recipients of machine support. (Page 20, Lines 790-803)
Comment 8: The implications lack clear recommendations. The study suggests AI is useful but does not address whether universities should regulate AI use or if students need training on ethical AI engagement.
Response 8: We have added a new paragraph addressing this concern, as seen below:
Further, given the widespread integration of GenAI tools among English major students, universities must take an active role in establishing clear guidelines for responsible and ethical use. There need to be policies drafted and implemented to manage the utilization of AI within academic work so that scholarly integrity is paired with innovation. Besides, effectively designed training programs need to be provided to students to enhance their digital literacy, and the focus should be on ethical engagement, critical evaluation of AI-produced content, and the risks of overreliance. These programs can equip students with the competencies essential to use AI-supported learning in a responsible and independent manner. Collaboration among student affairs offices, IT support services, and academic departments is needed to ensure that GenAI integration is aligned with institutional values and promotes academic excellence and autonomy among learners. (Pages 21-22, Lines 855-865)
Comment 9: The limitations section is insufficient. It mentions reliance on self-reported data but does not discuss response bias, potential overestimation of AI effectiveness, or institutional factors affecting adoption.
Response 9: We have revised the limitations incorporating your suggestions and reviewer 4’s suggestions, as seen below:
Generative AI technologies are becoming breakthrough cognitive co-pilots for Indo-nesian English major students, especially in facilitating academic writing, language ac-quisition, and creative tasks, consistent with global trends in AI-facilitated learning. Re-sults indicate deep acceptance and satisfaction; yet, the use of self-report measures re-stricts the capacity to capture various behavior and the long-term impact of AI use on the accomplishment of learning. In addition, despite the fact that the participants were re-cruited from a representative sample of institutions in Indonesia, the research is con-text-specific, and the cultural and institutional uniqueness of the Indonesian context can restrict the transferability of the results to other education systems. This includes the requirement of context-sensitive explanation and follow-up research in different settings. Subsequent studies employing longitudinal and observational study designs would better tell us about the long-term impact of generative AI tools on academic achievement and skill attainment. Intentional curriculum inclusion of GenAI calls for systematic training for appropriate use, critical thinking, and judicious dependence on technology. Having strong regulations in place and creating digital literacy will be crucial for making AI assets agents of academic excellence and innovation and reducing risks like over-dependence and ethics. (Page 21, Lines 838-853)
Reviewer 2 Report
Comments and Suggestions for Authors
This manuscript examines the utilization, user satisfaction and acceptance parameters associated with GenAI tools among students of the English major in Indonesian tertiary education, deploying a mixed-method research design. The work carried out is of commendable quality, meriting broader dissemination and subsequent research pioneering in its wake. However, there remain a handful of points that could be refined to fully realize the manuscript's potential.
- Abstract: The Abstract presents a remarkably lucid overview of the paper's breadth. Nonetheless, it would be advisable to refrain from numerically characterising statistical values and opt for expressing them more generally.
- Keywords: It is of utmost importance that the Keywords section avoids replicating the keywords featured in the manuscript's title. It is recommended that the existing keywords be reassessed and any redundancy corrected.
- Introduction: The study's justification is well-constructed within the Introduction and at the literature review's conclusion.
- Literature Review: The Literature Review delivers informative insights and centres on key areas pertaining to the research question, incorporating contemporary references from high-value sources.
- Methods: The Methods' sections provide comprehensive details concerning context, participants, instruments and procedures. However, it is advised to revise P8L315 given the reference to a non-existent Table 4.
- Results: The Results section is overflowing with redundant repetition of numerical findings detailed in the tables, complicating lucid identification of key findings at a swift glance. Summarizing the seminal findings in each segment without replicating the figures and percentages already presented in the related tables would be beneficial.
- Discussion and Implication: While comparisons are drawn with pre-existing studies, a more meticulous elucidation is required. The theoretical and practical implications should be explicitly articulated.
- Conclusion: Both the summary and study's limitations are distinctly depicted. However, the contribution of the study to the field's knowledge corpus is yet to be addressed.
- References: A comprehensive array of references is provided in the corresponding section.
Author Response
Responses for Reviewer 2
Comment 1: Abstract: The Abstract presents a remarkably lucid overview of the paper's breadth. Nonetheless, it would be advisable to refrain from numerically characterising statistical values and opt for expressing them more generally.
Response 1: Thank you very much for your suggestions. We have removed all the statistical results from the abstract. (Page 1, Lines 10-16)
Comment 2: Keywords: It is of utmost importance that the Keywords section avoids replicating the keywords featured in the manuscript's title. It is recommended that the existing keywords be reassessed and any redundancy corrected.
Response 2: We have revised all the keywords, as seen below:
Keywords: Generative AI assisted English learning; Cross-cultural technology adoption; Digital pedagogy; Language teaching innovation (Page 1, Lines 22-24)
Comment 3: Introduction: The study's justification is well-constructed within the Introduction and at the literature review's conclusion.
Response 3: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Comment 4: Literature Review: The Literature Review delivers informative insights and centres on key areas pertaining to the research question, incorporating contemporary references from high-value sources.
Response 4: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Comment 5: Methods: The Methods' sections provide comprehensive details concerning context, participants, instruments and procedures. However, it is advised to revise P8L315 given the reference to a non-existent Table 4.
Response 5: Thank you very much for the constructive feedback. We have remove “Table 4”.
Comment 6: Results: The Results section is overflowing with redundant repetition of numerical findings detailed in the tables, complicating lucid identification of key findings at a swift glance. Summarizing the seminal findings in each segment without replicating the figures and percentages already presented in the related tables would be beneficial.
Response 6: Thank you for your valuable feedback. We acknowledge the concern regarding repetition in the Results section. However, we chose to present selected numerical findings alongside the tables to ensure clarity and accessibility for readers who may focus on the narrative rather than the visual data. This approach also allows us to emphasize and interpret specific patterns and contrasts directly within the text, particularly for readers viewing the manuscript in print or without immediate access to the figures. That said, we have reviewed the Results section to ensure that all numerical descriptions serve an interpretive function and do not merely duplicate table content.
Comment 7: Discussion and Implication: While comparisons are drawn with pre-existing studies, a more meticulous elucidation is required. The theoretical and practical implications should be explicitly articulated.
Response 7: We have added both implications, as seen below:
As for the theoretical implications, this study extends the UTAUT model by contextu-alizing its constructs—particularly performance expectancy and effort expectan-cy—within the domain of English language learning in a Southeast Asian context. The findings affirm the predictive strength of performance expectancy, in line with Strzelecki (2024) and Habibi et al. (2023), but also underscore the unique cultural dynamics of Indo-nesian learners, where social influence, though modest, still reflects collectivist values. Moreover, students’ reflective engagement with AI aligns with cognitive engagement the-ory (Greene & Miller, 1996) and the ICAP framework (Chi & Wylie, 2014), emphasizing the pedagogical potential of AI tools as catalysts for constructive and interactive learning. The use of AI for idea generation, content exploration, and revision suggests a model of AI not just as a tool, but as a dynamic agent in self-regulated and metacognitive learning.
Practically, these findings have direct implications for higher education institutions seeking to integrate generative AI into language curricula. First, the study highlights the need for structured training programs that help students critically evaluate AI output and use tools ethically. Second, institutions should consider adopting flexible AI integration policies that support both innovation and academic integrity. Faculty development pro-grams should also be introduced to ensure instructors are equipped to guide students in reflective and responsible AI use. Lastly, AI tools should be positioned not as replacements for human instruction, but as pedagogical partners that complement traditional methods, supporting differentiated and inclusive learning experiences. (Page 21, Lines 818-836)
Comment 8: Conclusion: Both the summary and study's limitations are distinctly depicted. However, the contribution of the study to the field's knowledge corpus is yet to be addressed.
Response 8: We have added a new paragraph addressing this concern, as seen below:
This research adds to the expanding body of work on generative AI in education by utilizing a context-specific uptake and perceived usefulness analysis within Indonesian English major undergraduates—a context lacking in the prevailing body of work. By integrating quantitative and qualitative results into the UTAUT framework and cognitive engagement theories, this research expands earlier research on GenAI from Western-oriented contexts to Southeast Asia. It is discovering demographic nuances, task-domain usage, and reflective engagement patterns that are feeding into theory and practice. By so doing, the research is expanding our existing understanding of how students in multilingual, low-resource contexts interact with GenAI tools not just as technological devices, but as cognitive co-pilots impacting scholarly conduct, motivation, and agency. (Page 22, Lines 866-876)
Comment 9: References: A comprehensive array of references is provided in the corresponding section.
Response 9: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Reviewer 3 Report
Comments and Suggestions for Authors
Dear authors,
- line 39 I think there is a disconnect between the context and the preceding sentence. a relationship clause should be added between the sentences.
- Between lines 44 and 46, it is stated that demographic characteristics are not sufficiently taken into account. why should they be taken into account? after answering this question, I suggest that you indicate this gap in the literature. also, the contexts are disconnected from the preceding sections.
- what is your research problem? what problem situation did you think of in this article? what problem does this research involve the solution of? the problem situation is one of the most important parts of an article, so it should be very well structured.
-How did you prepare the questions between lines 76 and 85? There is no information.
- The literature review section is structured with current studies.
- heading 2.5. is poorly structured. the situation should be clarified with more grounded statements.
- The method section is well structured.
- 3.1. why is the heading numbered 3.1. under the title of method. shouldn't this be before the method?
- The analysis part of the article is also adequately structured.
- The findings section of the article is adequately structured and the discussion section is structured using current studies.
Author Response
Responses for Reviewer 3
Comment 1: - line 39 I think there is a disconnect between the context and the preceding sentence. a relationship clause should be added between the sentences.
Response 1: Thank you very much for your constructive feedback. We have revised it, as seen below:
Empirical studies further support these findings; a study analyzing qualitative responses from university student surveys in Sweden (Ou et al., 2024) and a mixed-method inquiry in Thailand revealed that students generally have positive perceptions of AI tools, viewing them as beneficial for language development and academic performance (Waluyo & Kusumastuti, 2024). (Pages 1-2, Lines 39-43)
Comment 2: - Between lines 44 and 46, it is stated that demographic characteristics are not sufficiently taken into account. why should they be taken into account? after answering this question, I suggest that you indicate this gap in the literature. also, the contexts are disconnected from the preceding sections.
Response 2: We have revised the whole paragraph, as seen below:
Although the integration of GenAI in education continues to grow, existing literature rarely addresses how demographic factors such as gender and academic level influence students' adoption and use of these tools, particularly in English language learning. Demographic characteristics are important to consider because they shape individual experiences, perceptions, and motivations toward educational technology. For example, despite widespread awareness of GenAI applications, female students and humanities majors tend to exhibit more negative attitudes toward them (Stöhr et al., 2024). Gender-specific priorities in technology adoption further highlight this divergence, with male students favoring compatibility, ease of use, and observability, while female students emphasize ease of use, compatibility, relative advantage, and trialability (Raman et al., 2024). Academic level and field of study also impact AI receptiveness, with technology and engineering students generally displaying more positive perspectives (Stöhr et al., 2024). However, targeted analysis of how these factors influence generative AI adoption in language learning remains limited, restricting insights into how varying acceptance patterns—such as those between genders or class levels—could inform more tailored support and resource allocation (Zhang et al., 2023). This oversight represents a critical gap in the literature, calling for more nuanced investigations that link technology adoption to learner diversity and educational equity. The present study seeks to fill this gap by examining how gender and academic year shape students’ engagement with GenAI tools in Indonesian English language education, offering demographic-specific insights that can inform more inclusive and targeted digital pedagogy. (Page 2, Lines 44-64)
Comment 3: - what is your research problem? what problem situation did you think of in this article? what problem does this research involve the solution of? the problem situation is one of the most important parts of an article, so it should be very well structured.
Response 3: To address this concern, we have revised the second paragraph and added one new paragraph as per suggestions (comments 2 and 4). Please see the changes we have made for comments 2 and 4. Also, please see our revision for your comment on No. 6.
Comment 4: -How did you prepare the questions between lines 76 and 85? There is no information.
Response 4: We have added a new paragraph, as seen below:
The research questions guiding this study were developed through a comprehensive synthesis of prior empirical and theoretical work on generative AI in education, particularly the Unified Theory of Acceptance and Use of Technology (Venkatesh et al., 2003) and its extensions in recent GenAI studies (e.g., Habibi et al., 2023; Jang, 2024; Waluyo & Kusumastuti, 2024). They were designed to explore not only the frequency and type of GenAI usage, but also learners’ satisfaction, perceived utility, and the sociotechnical factors influencing acceptance. Moreover, the final question reflects a shift in conceptual framing from mere usage to cognitive partnership, drawing on the notion of AI as a “co-pilot” in language learning (Chan & Hu, 2023; Du & Alm, 2024). A detailed synthesis of these empirical and conceptual foundations is provided in the following literature review section, which establishes the rationale and theoretical grounding for each research question. The combination of usage patterns, task types, satisfaction, adoption factors, and reflective learning practice ensures a comprehensive and context-sensitive inquiry into GenAI’s role in Indonesian higher education. (Page 3, Lines 101-114)
Comment 5: - The literature review section is structured with current studies.
Response 5: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Comment 6: - heading 2.5. is poorly structured. the situation should be clarified with more grounded statements.
Response 6: We have revised the paragraph, as seen below:
Although generative AI has been increasingly adopted in education for tasks such as writing, speaking, and reading (Aryadoust et al., 2024; Fathi et al., 2024; Wang, 2024), few studies focus specifically on its use within English language learning, especially in multilingual, non-Western contexts such as Indonesia. The literature often overlooks how AI tools are applied to particular academic tasks and how satisfaction with these tools relates to tangible educational outcomes (Barrett, 2023; Chan & Hu, 2023; Waluyo & Kusumastuti, 2024). Moreover, demographic factors such as gender and academic seniority are underexplored in GenAI adoption research, despite evidence showing that these variables shape attitudes and experiences with technology (Stöhr et al., 2024; Zhang et al., 2023; Rosmayanti et al., 2022).
Furthermore, while acceptance frameworks such as UTAUT have been widely used to assess GenAI adoption in higher education (Venkatesh et al., 2003; Habibi et al., 2023), their application to English major students remains limited. Most studies have focused on general populations or STEM disciplines, neglecting the pedagogical and cultural nuances of language education. In Southeast Asia, factors such as institutional support, collectivist learning norms, and unequal access to resources complicate the adoption process (Waluyo & Kusumastuti, 2024; Vo & Nguyen, 2024; Danler et al., 2024). Given these gaps, this study offers a context-specific investigation of GenAI use among English major students in Indonesian higher education, addressing tool usage, academic tasks, satisfaction, acceptance factors, and students' reflective engagement with AI as cognitive co-pilots. (Pages 6-7, Lines 280-299)
Comment 7: - The method section is well structured.
Response 7: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Comment 8: - 3.1. why is the heading numbered 3.1. under the title of method. shouldn't this be before the method?
Response 8: Thank you for your comment. We intentionally placed Section 3.1 under the Methods section to provide contextual grounding directly linked to the study’s research design. This section explains the linguistic and institutional challenges faced by English major students in Indonesia, which influenced key methodological decisions, including variable selection, instrument design, and interpretation of findings. Presenting this context here aligns with mixed-methods conventions, where setting-specific challenges are often integrated into the methods to justify the study's empirical focus (Creswell, 1999). We are happy to revise the structure if the editor deems it necessary but hope the rationale for its current placement is now clear.
Comment 9: - The analysis part of the article is also adequately structured.
Response 9: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Comment 10: - The findings section of the article is adequately structured and the discussion section is structured using current studies.
Response 10: Thank you very much for the constructive feedback. No revision was made based on this comment, but we did some revisions based on the reviewer’s comments.
Reviewer 4 Report
Comments and Suggestions for Authors
I would like to thank the authors for their revisions and responses. Please find further suggestions on areas of improvement below:
- Several sentences and claims in the article can cause confusion and need specificity.
- The study seems to be contextualized to one institution in one country. Overgeneralizations may not be appropriate.
- Several studies outside of North America have studied this topic. The authors need to make a deeper review of the literature.
- The study starts off by focusing on the demographics but then the results focus on the emergent trends of use.
- Numerous sentences are structured such that they are confusing or unclear. Major writing revisions are needed
- The study does not provide a clear and in-depth connection between the problem’s context and methodology. The rationale and justification, and application for the use of the approach are not enough
- The background references and literature review are broad and not well focused on the problem under study
- At times, it may seem that too many broad concepts are attempted to be brought and linked together in the article
- A write-up of results and discussion, well-grounded in visuals and tables, is missing
Author Response
Responses for Reviewer 4
Comment 1: Several sentences and claims in the article can cause confusion and need specificity.
Response 1: Thank you for pointing out the need for greater clarity and specificity in several parts of the manuscript. In response, we have carefully revised numerous sentences across the Introduction, Literature Review, Methods, and Discussion sections to eliminate ambiguity and enhance precision. These revisions include refining vague claims, clarifying terminology, and supporting statements with more specific examples or citations where appropriate. All changes are highlighted in blue in the revised manuscript for your review.
Comment 2: The study seems to be contextualized to one institution in one country. Overgeneralizations may not be appropriate.
Response 2: We have added it as one of the limitations, as seen below:
Generative AI technologies are becoming breakthrough cognitive co-pilots for Indonesian English major students, especially in facilitating academic writing, language acquisition, and creative tasks, consistent with global trends in AI-facilitated learning. Results indicate deep acceptance and satisfaction; yet, the use of self-report measures restricts the capacity to capture nuanced behavior and the long-term impact of AI use on the accomplishment of learning. In addition, despite the fact that the participants were recruited from a representative sample of institutions in Indonesia, the research is context-specific, and the cultural and institutional uniqueness of the Indonesian context can restrict the transferability of the results to other education systems. This includes the requirement of context-sensitive explanation and follow-up research in different settings. Subsequent studies employing longitudinal and observational study designs would better tell us about the long-term impact of generative AI tools on academic achievement and skill attainment. Intentional curriculum inclusion of GenAI calls for systematic training for appropriate use, critical thinking, and judicious dependence on technology. Having strong regulations in place and creating digital literacy will be crucial for making AI assets agents of academic excellence and innovation and reducing risks like over-dependence and ethics. (Page 21, Lines 838-853)
Comment 3: Several studies outside of North America have studied this topic. The authors need to make a deeper review of the literature.
Response 3: Thank you for your observation regarding the need to engage more deeply with studies conducted outside of North America. In response, we have revised the opening paragraph of the Literature Review section to explicitly highlight the breadth of international scholarship informing this study. This revised paragraph now foregrounds relevant research from diverse regions, including Asia (e.g., China, Iran, Indonesia, South Korea, Thailand, Vietnam), Europe (e.g., Poland, Spain, Sweden), and Oceania (e.g., New Zealand, Hong Kong). We believe this revision clarifies the global scope of the literature reviewed and positions our study within a broader international context. Please see the revised paragraph at the beginning of Section 2 (Literature Review), as seen below:
The worldwide spread of generative AI In education has produced burgeoning literature on its effects on language acquisition. Early scholarship was largely North American in origin, but recent scholarship has been drawn increasingly from a wide range of worldwide settings, providing rich descriptions of how generative AI technology is utilized across multiple cultural and institutional environments. For example, empirical studies on Iran (Fathi et al., 2024), China (An et al., 2023; Wang & Zhang, 2023; Xu et al., 2024), Vietnam (Vo & Nguyen, 2024), Thailand (Waluyo & Kusumastuti, 2024), Indonesia (Habibi et al., 2023; Rosmayanti et al., 2022), South Korea (Jang, 2024), and Hong Kong (Chan & Hu, 2023) explored students’ use of AI tools in EFL contexts. Evidence in New Zealand (Du & Alm, 2024), Sweden (Ou et al., 2024), and Poland (Strzelecki, 2024; Belda-Medina & Calvo-Ferrer, 2022) also confirms the extent of research on GenAI adoption and student attitudes within higher education. Such expanding international literature therefore provides necessary background to the analysis of how institutions and sociocultural factors shape student use behavior, attitudes, and learning outcomes. This current study rests on these myriad views, placing its research in the relatively underexplored field of Indonesian English language instruction. (Page 3, Lines 116-131)
Comment 4: The study starts off by focusing on the demographics but then the results focus on the emergent trends of use.
Response 4: We have addressed this concern by adding the following explanations in the introduction section:
Nevertheless, it is important to note that although the study is attentive to demographic factors such as gender and academic year, these variables are not examined in isolation. Rather, they serve as important lenses for interpreting emergent usage trends, satisfaction, and acceptance of GenAI tools. The primary emphasis of the analysis is on understanding the functional and reflective use of AI technologies in academic contexts, and how these patterns are influenced—but not wholly determined—by demographic variation. This framing allows the study to offer both generalizable insights into GenAI-supported learning and context-specific implications tied to learner diversity. (Page 2, Lines 78-86)
Comment 5: Numerous sentences are structured such that they are confusing or unclear. Major writing revisions are needed
Response 5: The revised manuscript has been thoroughly proofread and rewritten to improve clarity, eliminate ambiguity, and enhance overall readability.
Comment 6: The study does not provide a clear and in-depth connection between the problem’s context and methodology. The rationale and justification, and application for the use of the approach are not enough
Response 6: We have revised the method section which also followed the other reviewer's suggestions.
Comment 7: The background references and literature review are broad and not well focused on the problem under study
Response 7: We have extensively revised the introduction and literature review sections following your and other reviewers’ suggestions.
Comment 8: At times, it may seem that too many broad concepts are attempted to be brought and linked together in the article
Response 8: We have thoroughly revised the manuscript following the four reviewers’ suggestions. We hope that all the revisions have addressed this concern.
Comment 9: A write-up of results and discussion, well-grounded in visuals and tables, is missing
Response: We respectfully note that the original manuscript under review already included 2 charts, 1 figure, and 6 tables in the Results section, so we are unsure why the write-up was perceived as missing.
Round 2
Reviewer 1 Report
Comments and Suggestions for Authors
Thank you for addressing all my comments. I appreciate your hard work. I have no more comments. Just try to proofread your paper as there are some minor language issues.
Reviewer 3 Report
Comments and Suggestions for Authors
Dear authors,
According to the referee reports, the revisions you have made have improved the quality of your article.
Comments on the Quality of English Language
The English language of the article is sufficient.
Reviewer 4 Report
Comments and Suggestions for Authors
The research design and descriptions are not well outlined as seen in an empirical research paper. The methods need to follow a systematic approach that is exactly reproducible and leads to similar results. It is unclear if the research can contribute to the context since the frameworks adopted were not well situated and validated in the context of the country of interest. The novelty of the findings and the difference with the extensive body of literature on this topic is also unclear.