Next Article in Journal
Developing Inclusive Preschool Education for Children with Autism Applying Universal Learning Design Strategy
Previous Article in Journal
The Study of Core Practices in Support of More Ambitious Teacher Training: A Systematic Review (2019–2023)
Previous Article in Special Issue
Exploring ChatGPT’s Role in Higher Education: Perspectives from Pakistani University Students on Academic Integrity and Ethical Challenges
 
 
Article
Peer-Review Record

Artificial Intelligence in Higher Education: Bridging or Widening the Gap for Diverse Student Populations?

Educ. Sci. 2025, 15(5), 637; https://doi.org/10.3390/educsci15050637
by Dorit Hadar Shoval
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3:
Educ. Sci. 2025, 15(5), 637; https://doi.org/10.3390/educsci15050637
Submission received: 30 March 2025 / Revised: 19 May 2025 / Accepted: 19 May 2025 / Published: 21 May 2025

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

I have read the article, and I can say that it provides a valuable contribution to understanding the impact of AI in higher education. The text is well-written, employing a clear and concise academic style. The article is relevant and original, addressing a timely topic and highlighting a critical issue—the AI literacy divide, which is insufficiently discussed in the existing literature. However, a few aspects could be improved:

  1. The Abstract is well-structured, but it would benefit from a brief clarification of the methodology (e.g., how participants were selected, details about the research instruments used, and how the data were analyzed).
  2. Introduction - The concept of the "AI literacy divide" needs further clarification. While the effects on disadvantaged students are mentioned, both sides of the divide should be addressed for balance. Additionally, the study’s relationship with previous research could be more clearly stated.
  3. The methodology is well-structured but could benefit from further clarification for transparency and precision:

- A sentence explaining the study's innovation (e.g., using AI to reduce educational gaps) would strengthen this section.

- More detail is needed on the selection criteria for the 20 interviewed students, particularly regarding the balance in variables like socio-economic status, gender, and ethnicity. This would enhance the sample's validity.

- It would be useful to explain how the reflective journal was analyzed—how were relevant patterns and themes identified?

- More detail on the steps of thematic coding would help us understand how data was processed.

- A more detailed description of the data triangulation method would clarify how results were validated. Additionally, explaining how qualitative and quantitative data were integrated would improve interpretive clarity.

  1. Results - While quantitative results (significant group differences, frequency differences) are presented, there is insufficient methodological detail regarding the specific statistical analyses used. Detailing these analyses is crucial for demonstrating the validity and reliability of the results.

A paragraph explaining why certain statistical tests were chosen, along with their limitations (e.g., if the data do not meet the assumptions for parametric tests, this should be noted), should be added.

It would be helpful to clarify whether the results can be applied to other educational contexts or if they are specific to the studied context. For example, to what extent can these conclusions be extrapolated to other institutions or students in different academic fields?

  1. Conclusions and Discussion - Although the study's limitations are discussed in section 4.4, they could be more directly linked to the discussion results. For instance, the impact of sample size and the specifics of long-term studies on AI engagement and cognitive flexibility could be explored in greater detail. Additionally, self-reported student responses, another limitation, should be more thoroughly addressed in terms of how they may affect the interpretation of results related to AI usage perceptions and overall satisfaction. The conclusions section provides solid recommendations for educational institutions, but more suggestions could be added for professors and administrators on how to encourage more active student engagement with AI, considering the observed technological skill and cultural capital disparities. I also noticed that the article does not include sections such as: Author Contributions, Institutional Review Board Statement, Informed Consent Statement, Data Availability Statement, and Conflict of Interest.

Best regards

Author Response

Comment 1: The Abstract is well-structured, but it would benefit from a brief clarification of the methodology (e.g., how participants were selected, details about the research instruments used, and how the data were analyzed).

Response: Thank you for your valuable feedback regarding the abstract. I appreciate your suggestion to include a brief clarification of the methodology. I have revised the abstract to incorporate more details about the research instruments used (surveys, in-depth interviews, and the lecturer's reflective journal), participant numbers for quantitative and qualitative data collection, and a brief mention of the data analysis techniques employed (mixed-methods, statistical tests, thematic analysis, and triangulation). I believe this addition enhances the clarity and transparency of the study as presented in the abstract, in line with your recommendation. (line 4-20)

Comment 2: Introduction - The concept of the "AI literacy divide" needs further clarification. While the effects on disadvantaged students are mentioned, both sides of the divide should be addressed for balance. Additionally, the study’s relationship with previous research could be more clearly stated.

Response: Thank you for your insightful comments regarding the Introduction section. I appreciate your suggestions to further clarify the concept of the "AI literacy divide," ensure a balanced perspective on both sides of this divide, and better articulate the study's relationship with previous research. In response to your feedback, I have made the following revisions to the Introduction:

  • Clarification of "AI Literacy Divide": I have added a more explicit definition of the "AI literacy divide" where it is first introduced in Section 1.2 (lines 118-130). This definition now clarifies that the divide encompasses not only technical skills but also the cognitive and strategic abilities needed for effective AI utilization in academic contexts.
  • Balanced Perspective on the Divide: I have revised the text in Section 1.2 to provide a more balanced perspective on the widening educational gap (lines 118-130). I now explicitly state that while students with less technological proficiency may under-benefit, students who are already technologically adept or have greater access to AI tools may fully benefit and gain further advantages, thereby potentially widening the gap between these groups.
  • Relationship with Previous Research: I have added a sentence before the research questions in Section 1.2 to more clearly articulate how the present study builds upon existing theoretical frameworks of cultural and technological capital and the digital divide to address a critical gap in empirical research on the differential impact of AI on diverse student populations (lines 137-147).

I believe these revisions strengthen the Introduction by providing clearer definitions, a more balanced discussion of the "AI literacy divide," and a more explicit positioning of my study within the existing literature, as you suggested.

Comment 3: The methodology is well-structured but could benefit from further clarification for transparency and precision: A sentence explaining the study's innovation (e.g., using AI to reduce educational gaps) would strengthen this section. More detail is needed on the selection criteria for the 20 interviewed students, particularly regarding the balance in variables like socio-economic status, gender, and ethnicity. This would enhance the sample's validity. It would be useful to explain how the reflective journal was analyzed—how were relevant patterns and themes identified? More detail on the steps of thematic coding would help us understand how data was processed. A more detailed description of the data triangulation method would clarify how results were validated. Additionally, explaining how qualitative and quantitative data were integrated would improve interpretive clarity.

Response: Thank you for your constructive feedback on the Methodology section. I appreciate your suggestions for enhancing transparency and precision, particularly regarding the study's innovation, participant selection, and data analysis procedures. I have carefully considered each point and made the following revisions to improve this section:

  • Study's Innovation: I have added a sentence at the beginning of Section 2.1 (Research Approach) to explicitly state the innovative nature of this study, highlighting its focus on the differential impact of AI on diverse student populations within a real educational context (line 70).
  • Interviewee Selection Criteria: In Section 2.4 (Participants), I have added a sentence clarifying that the 20 interviewees were selected to reflect the demographic diversity of the course participants, aiming for representation across key variables such as gender, ethnic background (majority/minority group), first-generation status, and socioeconomic status as reported in the survey data (line 162-163).
  • Reflective Journal Analysis: In Section 2.6 (Data Analysis), within the Qualitative Analysis subsection, I have added details explaining that the lecturer's reflective journal was analyzed using the same thematic analysis method as the interviews (lines 293-310). I clarified that the journal provided systematic documentation of observations and reflections which were coded and analyzed to identify recurring patterns and themes relevant to student engagement and pedagogical responses.
  • Thematic Coding Steps: Also in Section 2.6, within the Qualitative Analysis subsection, I have provided more detail on the specific steps of thematic analysis performed, following Braun and Clarke's method (lines 293-310). This includes outlining the process from familiarization and initial code generation through theme development, review, definition, and the final report production.
  • Data Triangulation and Integration: In Section 2.6, under the Triangulation subsection, I have significantly expanded the description of my triangulation process (lines 311-320). I now explain how I compared and contrasted findings from the quantitative surveys, qualitative interviews, and the lecturer's journal to corroborate themes, explore discrepancies, and build a comprehensive understanding. I also clarified how the quantitative and qualitative data were integrated, noting how quantitative results informed the interpretation of qualitative data, and qualitative themes provided rich context for quantitative patterns, thus enhancing the robustness and depth of my findings.

I believe these additions provide the necessary detail and clarity to the Methodology section, addressing your valuable points and enhancing the transparency and rigor of my research description.

Comment 4: Results - While quantitative results (significant group differences, frequency differences) are presented, there is insufficient methodological detail regarding the specific statistical analyses used. Detailing these analyses is crucial for demonstrating the validity and reliability of the results. A paragraph explaining why certain statistical tests were chosen, along with their limitations (e.g., if the data do not meet the assumptions for parametric tests, this should be noted), should be added. It would be helpful to clarify whether the results can be applied to other educational contexts or if they are specific to the studied context. For example, to what extent can these conclusions be extrapolated to other institutions or students in different academic fields?

Response: Thank you for your feedback regarding the Results section. I appreciate your comments on the need for more methodological detail regarding the statistical analyses and clarification on the generalizability of the findings. I have addressed these points through the following:

  • Statistical Analysis Details: Regarding the methodological detail and rationale for the statistical tests, I have expanded Section 2.6 (Data Analysis) to provide a more detailed explanation of the specific statistical analyses used (independent samples t-tests and Chi-square tests), including the rationale for their selection based on the data type and research questions. I have also added a note about examining underlying assumptions where applicable. Effect sizes (Cohen's d and phi coefficient) have been added to Tables 2  and 3  and are mentioned in Section 2.6 (line 285-292).
  • Generalizability of Results: Regarding the generalizability of the findings to other educational contexts, this is an important point. As this study employed a case study approach focusing on a specific institution and student population, the generalizability is indeed limited. I have discussed this limitation extensively in Section 4.4 (Limitations) (lines 754-759), where I elaborate on how the specific context may influence the applicability of the results to other institutions, student populations, and academic fields. I believe that discussing this point in detail within the dedicated Limitations section is the most appropriate place to provide the necessary nuance and context.

I believe these adjustments enhance the clarity and robustness of the presentation of my findings.

Comment 5: Although the study's limitations are discussed in section 4.4, they could be more directly linked to the discussion results. For instance, the impact of sample size and the specifics of long-term studies on AI engagement and cognitive flexibility could be explored in greater detail. Additionally, self-reported student responses, another limitation, should be more thoroughly addressed in terms of how they may affect the interpretation of results related to AI usage perceptions and overall satisfaction. The conclusions section provides solid recommendations for educational institutions, but more suggestions could be added for professors and administrators on how to encourage more active student engagement with AI, considering the observed technological skill and cultural capital disparities. I also noticed that the article does not include sections such as: Author Contributions, Institutional Review Board Statement, Informed Consent Statement, Data Availability Statement, and Conflict of Interest.

Response: Thank you for your valuable final comments regarding the discussion, conclusions, and formal sections of the manuscript. I appreciate your detailed feedback and have addressed each point as follows:

  • Linking Limitations to Results and Discussion: I agree that explicitly linking the study's limitations to the discussion of the findings is crucial for a rigorous interpretation of the results. I have revised Section 4.4 (Limitations) (lines 753-788) to more directly connect each discussed limitation (including the single case study design (lines 757-759), study duration (lines 763-766), and the nature of self-reported data (lines 772-774) to its potential impact on the observed findings and the interpretation of concepts such as the AI literacy divide, AI engagement patterns, and AI-enhanced cognitive flexibility.
  • Enhanced Practical Recommendations: I appreciate your suggestion to provide more targeted recommendations for professors and administrators. In Section 5 (Conclusions), I have added more specific, actionable suggestions for educators and administrators on how to foster equitable AI engagement among diverse student populations (lines 814-823). These recommendations are grounded in my findings regarding the observed disparities in technological skills and cultural capital and include examples of concrete initiatives such as tiered training, integration of AI literacy, culturally responsive materials, and targeted support.
  • Formal Sections: Thank you for noting the absence of formal sections such as Author Contributions, Institutional Review Board Statement, Informed Consent Statement, Data Availability Statement, and Conflict of Interest within the main manuscript text. In accordance with the journal's submission guidelines, this information was provided separately through the online submission system during the initial submission process. This ensures adherence to the journal's standard procedure for handling these statements.

I believe these revisions, along with the previously discussed changes, significantly strengthen the manuscript and address your valuable feedback.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript addresses a timely and highly relevant topic: the integration of artificial intelligence (AI) in higher education and its potential to bridge or widen existing educational disparities among diverse student populations. While the study presents a robust research design and a compelling conceptual framework, several aspects require significant revision to meet the publication standards of Education Sciences. The abstract needs to be rewritten to conform to the IMRD format (Introduction, Methodology, Results, and Discussion). Currently, it lacks clear structure and reads more as a general summary than an academic abstract. The use of explicit signposting (e.g., “This study aimed to…”, “Using a mixed-methods approach…”, “Findings revealed…”, “These results suggest…”) is recommended to improve clarity and academic tone. Moreover, some sentences are overly long and should be revised for conciseness and precision. The introduction is extensive and well-documented, but overly descriptive at times. Key theoretical concepts—such as cultural capital, technological capital, and digital divide—are introduced sequentially but without sufficient critical integration. The section would benefit from a more cohesive articulation of the research gap, as well as a clearer positioning of the study within current debates on educational equity and AI. The research questions appear late in the text and in narrative form; it is recommended to explicitly label them as RQ1, RQ2, and RQ3 to improve clarity and facilitate later referencing. The methodology section presents a solid rationale for the use of a case study and mixed-methods approach. However, important details are missing. The triangulation process is not fully explained, and it remains unclear how integration between qualitative and quantitative data was achieved. It is also not specified whether more than one researcher was involved in coding the qualitative data, nor whether intercoder agreement procedures were followed. While the development of the AI prompt is described in detail, the implementation strategy and data collection instruments could be summarized more effectively. In addition, none of the tables follow APA formatting conventions. Formatting should be revised in all figures and tables to align with APA guidelines. The results are rich and well-organized around the research questions, but the section is overly long and, at times, repetitive. The qualitative findings would greatly benefit from the inclusion of a visual table summarizing subthemes alongside representative participant quotes (e.g., subtheme → example of quote). This would enhance readability and provide a clearer overview of the key findings. An example of such a presentation can be found in:

Manzano-León, A., Rodríguez-Rivera, P., Raposo-Rivas, M. et al. (2024). Digital game-based learning for transgender identity awareness: a qualitative study in Spanish social education degree. Current Psychology, 43, 18065–18073. https://doi.org/10.1007/s12144-023-05556-3

Quantitative results are well detailed and statistically robust, but the amount of numeric information may overwhelm readers. Tables 2 and 3 contain valuable comparative data but would benefit from clearer formatting and better alignment with APA style standards. The discussion section introduces a conceptual model that seeks to explain disparities in AI adoption and engagement across student groups. While this model is promising, the narrative is at times redundant with the results section. A deeper theoretical engagement is needed. For example, the concept of “AI-enhanced cognitive flexibility” is mentioned but not sufficiently developed or connected to existing literature. Likewise, “AI literacy divide” is introduced as a key construct but lacks a clear definition and operationalization. The discussion would benefit from a tighter conceptual thread that links the findings to broader debates in educational sociology, digital equity, and AI in learning environments.The conclusions summarize the study’s main insights but remain too general in their formulation. The recommendations for policy and practice are relevant but need to be more specific and directly grounded in the empirical data. A more critical reflection on the study’s limitations—especially regarding self-report bias, short study duration, and sample representativeness—is needed. The suggestion to conduct longitudinal and comparative studies in other institutional contexts is appropriate but could be elaborated with more detail.

Author Response

Comment 1: The abstract needs to be rewritten to conform to the IMRD format (Introduction, Methodology, Results, and Discussion). Currently, it lacks clear structure and reads more as a general summary than an academic abstract. The use of explicit signposting (e.g., “This study aimed to…”, “Using a mixed-methods approach…”, “Findings revealed…”, “These results suggest…”) is recommended to improve clarity and academic tone. Moreover, some sentences are overly long and should be revised for conciseness and precision.

Response: Thank you for your comprehensive comments, including your suggestions for improving the structure and clarity of the abstract. I agree that a clear, structured abstract is crucial. I have revised the abstract (lines 2-20) aiming to align it more closely with the IMRD structure you suggested (Background, Methods, Results, Conclusion) within a single paragraph. In adherence to the journal's specific guidelines which require a single paragraph abstract without headings, I have focused on improving the flow and implicit structure through careful phrasing and ordering of sentences, rather than using explicit signposting. I have also worked on refining the language for greater conciseness and academic tone. I hope the revised abstract provides a clearer and more structured overview of the study's key elements.

Comment 2: The introduction is extensive and well-documented, but overly descriptive at times. Key theoretical concepts—such as cultural capital, technological capital, and digital divide—are introduced sequentially but without sufficient critical integration. The section would benefit from a more cohesive articulation of the research gap, as well as a clearer positioning of the study within current debates on educational equity and AI. The research questions appear late in the text and in narrative form; it is recommended to explicitly label them as RQ1, RQ2, and RQ3 to improve clarity and facilitate later referencing.

Response: Thank you for your comprehensive feedback on the Introduction section. I appreciate your assessment of its scope and documentation, and I have carefully considered your suggestions for enhancing its analytical depth, theoretical integration, and clarity in positioning the study. In addressing your valuable comments, I have undertaken the following revisions to strengthen the Introduction:

  • Theoretical Integration and Analytical Depth: While I aimed for a thorough review of the relevant theoretical frameworks, I recognize the importance of a more explicit and critical integration of these concepts (cultural capital, technological capital, digital divide) in relation to the study's core focus on AI in higher education. I have revised sections 1.1 and 1.2 to strengthen the conceptual links between these established theories and the emerging dynamics of AI adoption and its potential impact on educational disparities. I have added phrasing to emphasize how the study utilizes these frameworks to provide a critical perspective on AI's role in shaping educational equity and to underscore the importance of understanding these dynamics for inclusive AI implementation strategies (lines 78-82).
  • Articulation of Research Gap and Study Positioning: I have refined the language in Section 1.2 (lines 137-147) to provide a more cohesive articulation of the research gap my study addresses, specifically, the lack of in-depth empirical research on the differential effects of AI on diverse student populations (lines 139-141). I have also added a sentence towards the end of Section 1.2 to more clearly position the research within the current critical debates on whether technological advancements in education serve to promote equity or reinforce existing social stratifications (line 147-149).
  • Research Question Labeling: To improve clarity and facilitate referencing throughout the manuscript as you suggested, I have explicitly labeled the main research questions as RQ1, RQ2, and RQ3 in Section 1.2 (lines 152-157).

I believe these revisions enhance the Introduction by fostering a more integrated discussion of the theoretical underpinnings and providing a clearer articulation of the study's contribution to the ongoing academic conversation on AI and educational equity. I aimed to incorporate your feedback while maintaining the intended scope and structure of the section.

Comment 3: The methodology section presents a solid rationale for the use of a case study and mixed-methods approach. However, important details are missing. The triangulation process is not fully explained, and it remains unclear how integration between qualitative and quantitative data was achieved. It is also not specified whether more than one researcher was involved in coding the qualitative data, nor whether intercoder agreement procedures were followed. While the development of the AI prompt is described in detail, the implementation strategy and data collection instruments could be summarized more effectively. In addition, none of the tables follow APA formatting conventions. Formatting should be revised in all figures and tables to align with APA guidelines.

Response: Thank you for your detailed and constructive comments regarding the Methodology section. I appreciate your acknowledgment of the study's research design and mixed-methods approach and your valuable suggestions for increasing the transparency and precision of its description. I have carefully addressed the points you raised:

  • Triangulation and Data Integration: I agree that a clear explanation of the triangulation process and data integration is essential. I have expanded Section 2.6 (Data Analysis), particularly within the Triangulation subsection (lines 311-320), to provide a more detailed description of how findings from the quantitative surveys, qualitative interviews, and the lecturer's journal were compared and integrated to corroborate results, explore divergent insights, and build a comprehensive understanding of the phenomenon. I explained how qualitative data provided rich context for quantitative patterns and how quantitative findings informed the interpretation of qualitative themes.
  • Qualitative Coding and Intercoder Agreement: You correctly noted the importance of detailing the qualitative coding process. I have added information to Section 2.6 (Data Analysis) under the Qualitative analysis subsection to clarify that two researchers were involved in the coding process (line 305-310). I report that a subset of the interview transcripts (25%) was independently coded by two researchers, and intercoder reliability was assessed using Cohen's Kappa, yielding a value of 0.85, indicating substantial agreement. This enhanced the trustworthiness of the qualitative analysis.
  • Implementation Strategy and Data Collection Instruments: The implementation strategy is detailed in Section 2.3 (Technical Implementation Details), describing the development of the simulation prompt and its classroom implementation. The data collection instruments are described in Section 2.5 (Research Instruments). To provide full access to the details of the primary instruments, the complete simulation prompt is presented as Figure 1 in the manuscript, and the in-depth interview guide is included in Appendix A. I believe that providing these detailed resources, alongside the descriptions in Sections 2.3 and 2.5, offers a comprehensive account of the research tools and their implementation.
  • Table and Figure Formatting: I appreciate you highlighting the importance of clear and consistent formatting for tables and figures. I have reviewed all tables (Tables 1, 2, and 3) and figures (Figure 1 and 2) to ensure they comply with the specific guidelines outlined in the journal's Instructions for Authors for preparing visual elements. I have made necessary adjustments to enhance their clarity and consistency in accordance with these guidelines and standard academic presentation formats. I will ensure strict adherence to the journal's requirements upon submission of the revised manuscript.

I believe these revisions and clarifications significantly enhance the transparency and rigor of the Methodology section, addressing your valuable feedback.

Comment 4: The results are rich and well-organized around the research questions, but the section is overly long and, at times, repetitive. The qualitative findings would greatly benefit from the inclusion of a visual table summarizing subthemes alongside representative participant quotes (e.g., subtheme → example of quote). This would enhance readability and provide a clearer overview of the key findings. An example of such a presentation can be found in: [Reference to example article]. Quantitative results are well detailed and statistically robust, but the amount of numeric information may overwhelm readers. Tables 2 and 3 contain valuable comparative data but would benefit from clearer formatting and better alignment with APA style standards.

Response: Thank you for your detailed assessment of the Results section. I appreciate your recognition of the richness of the findings and their organization around the research questions. I have carefully considered your valuable suggestions regarding the length, presentation of qualitative data, quantitative information, and table formatting. In response to your feedback:

  • Length and Repetitiveness: I acknowledge your observation that the section may appear long and at times repetitive. My approach to presenting the quantitative and qualitative findings for each research question, often side-by-side or sequentially (Section 3), was intentional. This structure was chosen to explicitly demonstrate the triangulation of data from different sources (surveys, interviews, lecturer's journal) around the research questions. I believe that showing how consistent themes and patterns emerged across multiple methods strengthens the validity and trustworthiness of my findings. While this presentation may lead to some overlap in the description of findings, it is crucial for illustrating the empirical basis of the identified themes and the robustness of the analysis through triangulation. I have, however, reviewed the section to minimize redundancy where possible without compromising this necessary detailed presentation.
  • Qualitative Findings Presentation: I appreciate your suggestion to include a visual table summarizing qualitative subthemes and quotes, and I reviewed the example you provided with interest. I agree that such a table can offer a concise overview. However, given my decision to maintain a rich, detailed narrative that integrates qualitative insights directly with quantitative findings for triangulation purposes, and to avoid adding further length to the manuscript, I have chosen to retain the current format. My qualitative findings are presented thematically with representative quotes integrated into the text, allowing for a detailed and contextualized understanding of the student experiences.
  • Quantitative Information: I understand that the amount of numeric information in the text may be overwhelming for some readers. The inclusion of specific statistical data (means, standard deviations, test statistics, p-values, and effect sizes) in the narrative is intended to highlight the key findings discussed and provide immediate empirical support for the points being made, in addition to the detailed information presented in the tables. I have reviewed the section to ensure that the narrative flow is maintained despite the inclusion of these data.
  • Table Formatting: I appreciate you pointing out the formatting of the tables. I have reviewed all tables (Tables 1, 2, and 3)  to ensure they comply with the specific guidelines for preparing figures, schemes, and tables as outlined in the journal's Instructions for Authors. I have made necessary adjustments to enhance their clarity and consistency in accordance with these guidelines and standard academic presentation formats. I will ensure strict adherence to the journal's requirements upon submission of the revised manuscript.

I believe this approach, while maintaining the detailed presentation crucial for demonstrating triangulation, addresses your comments by explaining the rationale behind the structure and ensuring compliance with formatting standards.

Comment 5: The discussion section introduces a conceptual model that seeks to explain disparities in AI adoption and engagement across student groups. While this model is promising, the narrative is at times redundant with the results section. A deeper theoretical engagement is needed. For example, the concept of “AI-enhanced cognitive flexibility” is mentioned but not sufficiently developed or connected to existing literature. Likewise, “AI literacy divide” is introduced as a key construct but lacks a clear definition and operationalization. The discussion would benefit from a tighter conceptual thread that links the findings to broader debates in educational sociology, digital equity, and AI in learning environments.

Response: Thank you for your valuable feedback on the Discussion section and for recognizing the potential of the conceptual model presented (Figure 2). I have carefully considered your suggestions for enhancing the section's clarity, theoretical depth, and conceptual coherence. In response to your comments:

  • Redundancy with Results: I acknowledge your observation that the discussion narrative at times overlaps with the results section. While some degree of overlap is inherent in integrating quantitative and qualitative findings and referencing them in the discussion, I have reviewed and revised the Discussion section to minimize redundancy where possible. My focus in the revised section is more explicitly on interpreting the findings through the lens of the proposed conceptual framework and theoretical constructs, rather than merely restating the results.
  • Deeper Theoretical Engagement, Definition, and Operationalization: I appreciate your point regarding the need for deeper theoretical engagement and clearer definition of key concepts. I have made revisions in Sections 4.1 to address this. I have added more explicit definitions for the core concepts of the model, such as the "AI literacy divide" (lines 644-648, 651-654) and "AI-enhanced cognitive flexibility" (lines 685-690, 692-695), clarifying how these concepts are understood and evidenced within the context of this specific study and its findings (effectively providing an operationalization within this framework). Furthermore, I have strengthened the explicit connections between these concepts and the foundational theoretical frameworks discussed in the Introduction (cultural capital, technological capital, digital divide) and relevant literature.
  • Tighter Conceptual Thread and Link to Broader Debates: To enhance the conceptual thread and better link the findings to broader academic discussions, I have worked on the transitions and connecting phrases within and between the subsections of the Discussion. I have aimed to more clearly articulate how the patterns observed in AI adoption and engagement among diverse student groups contribute to ongoing debates in educational sociology, digital equity, and the integration of AI in learning environments. I have also added a brief concluding paragraph before the limitations section to summarize the main contributions of the model and findings to these wider discussions (lines 743-751).

I believe these revisions improve the clarity, theoretical grounding, and coherence of the Discussion section, addressing your insightful comments and highlighting how my study's findings and conceptual model contribute to the relevant academic fields.

Comment 6: The conclusions summarize the study’s main insights but remain too general in their formulation. The recommendations for policy and practice are relevant but need to be more specific and directly grounded in the empirical data. A more critical reflection on the study’s limitations—especially regarding self-report bias, short study duration, and sample representativeness—is needed. The suggestion to conduct longitudinal and comparative studies in other institutional contexts is appropriate but could be elaborated with more detail.

Response: Thank you for your valuable feedback on the Conclusions section. I appreciate your suggestions for enhancing the specificity of the conclusions and recommendations, as well as the detail in the limitations and future research directions. I have carefully considered your comments and made the following revisions:

  • Specificity of Conclusions: I have revised the introductory part of Section 5 to make the summary of the study's main insights more specific. The conclusions now explicitly link the findings to the proposed conceptual framework and highlight key empirical results, such as the emerging AI literacy divide and differential engagement patterns, providing a more concrete overview of the study's contributions. (line 790-799).
  • Specificity and Data-Grounding of Recommendations: I have reviewed the recommendations for policy and practice in Section 5. These recommendations, particularly those directed at educators and administrators, were previously expanded upon. I believe they are now sufficiently specific and grounded in the empirical data by addressing the observed disparities and differential engagement patterns highlighted in the results (line 814-823). For example, the recommendations for tiered training and culturally responsive materials directly respond to the findings on varying technological exposure and cultural/linguistic factors influencing AI engagement. I have also added a paragraph explicitly grounding the recommendations in the ethical imperative to ensure educational equity (lines 824-828).
  • Critical Reflection on Limitations: Regarding the critical reflection on the study's limitations (such as self-report bias, study duration, and sample characteristics), these are indeed important considerations. I have addressed the study's limitations in detail in Section 4.4 (Limitations), where I discuss how specific limitations may impact the interpretation of the findings, providing the critical reflection you requested (line 753-788).
  • Elaboration of Future Research Suggestions: I have expanded the suggestions for future research in Section 5. I have added more detail or rationale for each suggested direction, linking them to the study's limitations and key findings, to provide a clearer roadmap for subsequent investigations into AI and educational equity (841-847).

I believe these revisions enhance the clarity and utility of the Conclusions section, addressing your valuable feedback while aligning with the overall structure and intent of the manuscript.

 

Reviewer 3 Report

Comments and Suggestions for Authors

At the beginning, I would like to acknowledge the team that developed the study Artificial Intelligence in Higher Education: Bridging or Widening the Gap for Diverse Student Populations? which I believe is a significant contribution to the field of integrating emerging technologies, in particular AI-based simulations, within higher education. The work contains a robust and well-structured approach to addressing a topic of relevance, especially in diverse and vulnerable educational contexts. With all due caution, I will review the different aspects of the paper and make, where appropriate, considerations that will serve to generate much-needed discussion on an emerging issue that has an impact on the education of the future.
The theoretical framework of the study is well grounded and adequately connects the key concepts with the central issue of the research, which is the impact of AI in higher education. The integration of theories such as Bourdieu's theories of cultural capital and Lee and Chen's theories of technological capital provides a solid conceptual basis for understanding how prior disparities can influence students' experience with technology. This theoretical approach allows contextualising the authors' observations on the gaps between different student groups (e.g. between first-generation students and others), which is crucial for understanding the results of the study. The concept of cognitive flexibility is also well articulated in the theoretical framework, aligning with contemporary discussions on how AI tools can enhance learning by fostering students' cognitive adaptability. The connection to van Dijk's theory of digital inequalities is justified by extending the idea of how technological inequalities affect not only access to technology, but also its use and the depth of interactions with digital tools.
A possible improvement would be a more explicit discussion of how key concepts such as technological capital or cognitive flexibility are defined and operationalised in the specific context of this study. Although they are mentioned in a general way, it could be useful to specify in more detail how these concepts are applied in the design of the educational intervention and in the analysis of the results. In addition, it would be relevant to include more recent references on the use of AI in educational settings, which could enrich the theoretical basis and provide a more updated frame of reference on emerging trends in technology-enhanced education.
The Materials and Methods section presents a clear and appropriate methodological approach to the study of the impact of AI-based simulations in education. The use of an observational design and the descriptive case study are well justified and allow to effectively address the phenomenon in a real and complex context. Furthermore, the choice of a mixed approach (quantitative and qualitative) and the triangulation of data strengthen the validity and completeness of the findings. The integration of generative AI-based simulations in a culturally sensitive way also deserves my consideration, reflecting an adequate consideration of the diversity of the student body. I highlight the clarity in the description of the methods, from the development of prompts to the analysis of the data, which allows the reader to understand how the research was conducted and what tools were used. Cultural and linguistic sensitivity is also a strength of the study, given the diversity of the educational context. Likewise, the rigour in the triangulation of data is a strength, as it ensures greater validity in the results obtained.

If you would like to take it into consideration, I recommend improving the section by clarifying some methodological details. In particular, it would be useful to specify the hypotheses of the research, if any, and to provide more information on the inclusion/exclusion criteria of the participants, especially regarding the selection of the interviewees. And I say clarify because I understand, and I may be wrong, that the implicit hypothesis would be: ‘The adoption of artificial intelligence (AI) tools in educational settings has the potential to enhance the learning experience and cognitive flexibility of students, but this implementation may also accentuate existing inequalities related to the technological and cultural capital of students, especially among those from historically marginalised groups, such as first-generation and minority students’. Furthermore, it would be relevant to clarify whether theoretical saturation was reached in the interviews, which could add robustness to the qualitative part of the analysis. These details would contribute to improving the transparency of the research process and the replicability of the study.
In the results section I report a rigorous and coherent presentation of the findings, clearly organised around the three research questions and with an effective integration of quantitative and qualitative data. I recognise the richness of the nuances provided by the student testimonies and the inclusion of the teachers' voice, which allows for a deeper understanding of the processes experienced. The triangulation of data strengthens the validity of the analysis and helps to identify differentiated patterns of technological appropriation according to variables of social, cultural and academic origin. However, I suggest that you take care to avoid redundancy in the presentation of results, especially at points where quantitative and qualitative data coincide; in these cases, a more refined synthesis could increase the communicative power of the text. It would also be useful to go deeper into the interpretation of some statistical contrasts, contextualising them in possible explanatory mechanisms beyond the mere observation of inequalities (for example, differences in linguistic capital, academic confidence, or institutional expectations). 
The discussion is, in my view, one of the strongest sections of the paper, clearly articulating the empirical findings with relevant theoretical frameworks such as cultural and technological capital and providing valuable concepts such as AI literacy and AI-enhanced cognitive flexibility. I underline the textual structure to highlight how emerging technologies can both offer opportunities and reinforce pre-existing inequalities, and how these dynamics are expressed in the differentiated usage patterns among student groups. However, it would be interesting to further deepen the theoretical discussion with a more critical problematisation of the limits of the categories used, as well as to expand the reflection on the ethical implications of the findings, especially in relation to the risk of consolidating a higher education stratified by access and technological capital - for which critical pedagogy would be the counterpoint and balance.
I would like to thank them for having developed work that offers a deep insight into how AI is shaping the educational landscape, especially in diverse contexts.

Author Response

Comment 1: The theoretical framework of the study is well grounded and adequately connects the key concepts with the central issue of the research, which is the impact of AI in higher education. The integration of theories such as Bourdieu's theories of cultural capital and Lee and Chen's theories of technological capital provides a solid conceptual basis for understanding how prior disparities can influence students' experience with technology. This theoretical approach allows contextualising the authors' observations on the gaps between different student groups (e.g. between first-generation students and others), which is crucial for understanding the results of the study. The concept of cognitive flexibility is also well articulated in the theoretical framework, aligning with contemporary discussions on how AI tools can enhance learning by fostering students' cognitive adaptability. The connection to van Dijk's theory of digital inequalities is justified by extending the idea of how technological inequalities affect not only access to technology, but also its use and the depth of interactions with digital tools. A possible improvement would be a more explicit discussion of how key concepts such as technological capital or cognitive flexibility are defined and operationalised in the specific context of this study. Although they are mentioned in a general way, it could be useful to specify in more detail how these concepts are applied in the design of the educational intervention and in the analysis of the results. In addition, it would be relevant to include more recent references on the use of AI in educational settings, which could enrich the theoretical basis and provide a more updated frame of reference on emerging trends in technology-enhanced education.

Response: Thank you very much for your thorough review and very positive assessment of my theoretical framework and Introduction section. I sincerely appreciate your recognition of the framework's solid grounding, its connection to the core research issue, and the effective integration of theories such as cultural capital, technological capital, and digital inequalities. Your comments highlighting the importance of this theoretical basis for contextualizing my findings are highly valued. I have carefully considered your suggestions for further improvement:

  • Explicit Definition and Operationalization of Key Concepts: You suggested a more explicit discussion of how concepts like technological capital and cognitive flexibility are defined and operationalized in the specific context of this study. This is a very insightful point. Concepts such as the "AI literacy divide"  and "AI-enhanced cognitive flexibility"  are integral parts of the novel conceptual framework that emerged from the data analysis conducted in this study. As such, their detailed definition, conceptualization, and operationalization within the context of the findings are presented and discussed in Section 4 (Discussion) (lines 605-751), where the framework is introduced in relation to the empirical results. Please see Section 4 for the in-depth discussion of these concepts as they are used in this study.
  • Inclusion of More Recent References: Thank you for suggesting the inclusion of more recent references on the use of AI in educational settings. I have reviewed the literature cited in the Introduction and throughout the manuscript and have aimed to include recent and relevant publications to provide an updated frame of reference. For example, the manuscript includes references to recent work by Dwivedi et al. (2023), Wu (2023), Cai et al. (2023), Hadar Shoval et al. (2024), O'Dea (2024), and Zhan et al. (2024), among others, reflecting current developments and discussions regarding AI in education.

I believe these clarifications, particularly regarding the location of the definition and operationalization of the emerging concepts, address your valuable suggestions while maintaining the appropriate structure for presenting a framework that emerged from my research findings.

Comment 2: The Materials and Methods section presents a clear and appropriate methodological approach to the study of the impact of AI-based simulations in education. The use of an observational design and the descriptive case study are well justified and allow to effectively address the phenomenon in a real and complex context. Furthermore, the choice of a mixed approach (quantitative and qualitative) and the triangulation of data strengthen the validity and completeness of the findings. The integration of generative AI-based simulations in a culturally sensitive way also deserves my consideration, reflecting an adequate consideration of the diversity of the student body. I highlight the clarity in the description of the methods, from the development of prompts to the analysis of the data, which allows the reader to understand how the research was conducted and what tools were used. Cultural and linguistic sensitivity is also a strength of the study, given the diversity of the educational context. Likewise, the rigour in the triangulation of data is a strength, as it ensures greater validity in the results obtained. If you would like to take it into consideration, I recommend improving the section by clarifying some methodological details. In particular, it would be useful to specify the hypotheses of the research, if any, and to provide more information on the inclusion/exclusion criteria of the participants, especially regarding the selection of the interviewees. Furthermore, it would be relevant to clarify whether theoretical saturation was reached in the interviews, which could add robustness to the qualitative part of the analysis. These details would contribute to improving the transparency of the research process and the replicability of the study.

Response: Thank you for your positive comments on the Methodology section and for recognizing the strengths of the mixed-methods case study approach, triangulation, and cultural sensitivity. I appreciate your suggestions for enhancing the transparency of certain methodological details. I have carefully considered your points and have made revisions or provided clarifications in the manuscript as described below:

  • Research Hypotheses: My study employs a descriptive case study approach (Section 2.1 - Research Approach) (lines 162-171) guided by specific research questions (Section 1.2) (lines 152-157), rather than formal hypotheses. This design is focused on providing a rich, contextualized understanding of a phenomenon within its real-life context, allowing for a more inductive process where the conceptual framework presented in the discussion emerged from the data analysis, which is typical of this type of mixed-methods case study.
  • Inclusion/Exclusion Criteria and Interviewee Selection: I agree that clarity on participant selection, especially for interviews, is important. In Section 2.4 (Participants) (lines 238-254), I have added more detail on the selection of the 20 interviewees (lines 252-254). I clarified that while all course participants were invited, those who agreed were selected with the explicit aim of reflecting the demographic diversity of the class reported in the survey data, aiming for representation across key variables such as gender, ethnic background, first-generation status, and socioeconomic status. Due to the voluntary nature of participation, strict inclusion/exclusion criteria were not applied beyond enrollment in the course and willingness to be interviewed.
  • Theoretical Saturation: Thank you for raising the point about theoretical saturation. While formal theoretical saturation was not the primary method of validation for the qualitative findings in this study, the robustness of the qualitative results was validated through triangulation across different data sources, as described in Section 2.6 (Data Analysis). The consistency of themes emerging from the interviews and the lecturer's journal, which were also supported by the quantitative findings showing significant group differences, provided strong evidence for the validity and reliability of the identified qualitative patterns.
  • General Methodological Details: As also requested by other reviewers, I have expanded the description of the data analysis procedures in Section 2.6 (lines 282-320). This includes providing more detail on the steps of thematic analysis used for the qualitative data (lines 293-310), specifying the statistical tests used for quantitative analysis (lines 285-292), and expanding the description of the triangulation process (lines 311-320).

I believe these revisions and explanations provide the necessary detail and transparency for the Methodology section, addressing your valuable points and enhancing the transparency and rigor of my research description.

Comment 3: In the results section I report a rigorous and coherent presentation of the findings, clearly organised around the three research questions and with an effective integration of quantitative and qualitative data. I recognise the richness of the nuances provided by the student testimonies and the inclusion of the teachers' voice, which allows for a deeper understanding of the processes experienced. The triangulation of data strengthens the validity of the analysis and helps to identify differentiated patterns of technological appropriation according to variables of social, cultural and academic origin. However, I suggest that you take care to avoid redundancy in the presentation of results, especially at points where quantitative and qualitative data coincide; in these cases, a more refined synthesis could increase the communicative power of the text. It would also be useful to go deeper into the interpretation of some statistical contrasts, contextualising them in possible explanatory mechanisms beyond the mere observation of inequalities (for example, differences in linguistic capital, academic confidence, or institutional expectations).

Response: Thank you again for your positive feedback on the Results section and your constructive suggestions.

  • Regarding the suggestion to avoid redundancy, particularly where quantitative and qualitative data coincide: My approach to presenting the quantitative and qualitative findings for each research question, often side-by-side or sequentially (Section 3), was intentional. This structure was chosen to explicitly demonstrate the triangulation of data from different sources (surveys, interviews, lecturer's journal) around the research questions. I believe that showing how consistent themes and patterns emerged across multiple methods strengthens the validity and trustworthiness of my findings. While this presentation may lead to some overlap in the description of findings, it is crucial for illustrating the empirical basis of the identified themes and the robustness of the analysis through triangulation. I have, however, reviewed the section to minimize redundancy where possible without compromising this necessary detailed presentation.
  • Concerning the suggestion to go deeper into the interpretation of some statistical contrasts and contextualize them within possible explanatory mechanisms (such as linguistic capital, academic confidence, or institutional expectations): A deeper interpretation and discussion of the findings, including the statistical contrasts, is primarily provided within the Discussion section (Section 4). In Section 4, I interpret the quantitative results (including the observed group differences) through the lens of the proposed conceptual framework and the theoretical constructs, drawing on the qualitative data to provide context and insight into potential factors underlying these differences (e.g., prior technological exposure, cultural background). While the mechanisms you suggest (linguistic capital, academic confidence, institutional expectations) are indeed very insightful and likely play a role in shaping student experiences with AI, exploring their specific influence as direct explanatory mechanisms for the statistical differences observed in this study's results would extend beyond the scope of the data I collected and analyzed in this particular research. My study identified the differences and proposed a framework based on the data; investigating the precise mediating role of factors like academic confidence or specific institutional expectations would require a different study design. I believe the Discussion section effectively interprets the findings within the bounds of the collected data, and the factors you mentioned represent important avenues for future research to explore these potential mechanisms more directly. I have also included effect sizes for the statistical tests in Tables 2 and 3 and mentioned them in Section 2.6 to provide additional context on the magnitude of the observed quantitative differences.

I trust that this clarifies the structure of the Results section and where the deeper interpretation is located within the manuscript, while also acknowledging the valuable directions for future research that your suggestion highlights.

Comment 3: The discussion is, in my view, one of the strongest sections of the paper, clearly articulating the empirical findings with relevant theoretical frameworks such as cultural and technological capital and providing valuable concepts such as AI literacy and AI-enhanced cognitive flexibility. I underline the textual structure to highlight how emerging technologies can both offer opportunities and reinforce pre-existing inequalities, and how these dynamics are expressed in the differentiated usage patterns among student groups. However, it would be interesting to further deepen the theoretical discussion with a more critical problematisation of the limits of the categories used, as well as to expand the reflection on the ethical implications of the findings, especially in relation to the risk of consolidating a higher education stratified by access and technological capital - for which critical pedagogy would be the counterpoint and balance.

Response: Thank you for your very positive assessment of the Discussion section and for providing constructive suggestions for further enhancement. I fully agree that a deeper critical engagement with the theoretical implications and an expanded reflection on the ethical dimensions are crucial. I have addressed these points in the revised manuscript:

  • Regarding the suggestion to further deepen the theoretical discussion with a more critical problematisation of the limits of the categories used: I have added a sentence in the introductory part of the Discussion section (Section 4) (lines 626-632) to introduce a more critical perspective on how AI integration challenges and reshapes existing theoretical frameworks and the concepts I use to understand inequality in the digital age. This addition acknowledges that established concepts like cultural and technological capital may manifest differently in the AI era and recognizes the context-dependent nature of emerging concepts like the AI literacy divide. This aims to set a slightly more critical tone for the discussion that follows.
  • Concerning the suggestion to expand the reflection on the ethical implications of the findings, especially in relation to the risk of stratification by access and technological capital, and the role of critical pedagogy: I agree that the risk of AI integration consolidating a stratified higher education system is a critical ethical concern that warrants explicit discussion. I have addressed this in the Conclusions section (Section 5) by adding a dedicated paragraph following the initial summary of findings (lines 800-806). This new paragraph explicitly discusses the significant ethical implications of the observed disparities (AI literacy divide, differential engagement) and highlights the urgent ethical responsibility of institutions to proactively address the risk of exacerbating inequalities and creating a more stratified system based on technological capital. Furthermore, within the section detailing practical recommendations in Section 5, I have added a sentence clarifying that these recommendations are grounded in this ethical imperative to ensure educational equity (lines 824-828) and are aimed at actively counteracting the observed tendencies towards increased stratification by empowering all students. While I did not explicitly use the term "critical pedagogy," the focus on challenging existing disadvantages, promoting equity, and empowering underserved students aligns strongly with its core principles and addresses the spirit of your suggestion regarding a necessary counterpoint to stratification.

I believe these additions in Section 4 and Section 5 significantly strengthen the discussion of the theoretical and ethical dimensions of the findings, directly addressing your valuable feedback.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

On the basis of the suggested corrections, I believe that the revisions address each comment precisely and serve to enhance the manuscript’s internal coherence and argumentative strength

Author Response

Thank you very much for your positive feedback on the revised manuscript. I am pleased that you found that the revisions addressed your comments precisely and served to enhance the manuscript's internal coherence and argumentative strength. Your initial constructive feedback was invaluable in guiding these improvements, and I sincerely appreciate your time and careful review.

Back to TopTop