The Impact of AI on Inclusivity in Higher Education: A Rapid Review

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article examines a timely and important topic – inclusivity in AI use in educational settings – and offers a comprehensive overview of AI applications in higher education. However, the manuscript requires extensive revision before it can be considered publishable.

 

The theoretical section needs to more clearly establish the research gap. At present, it is unclear how this study is conceptually and theoretically situated in relation to previous studies. The section reads more like a literature review than a theoretical framework, which is particularly confusing given that the methodology is a rapid review. This section would benefit from a revised title and clearer articulation of the research questions in relation to prior studies.

 

Furthermore, the analysis does not appear to emerge naturally from the review but instead feels predetermined and overly structured. One reason is that the general benefits of AI and the potential benefits for inclusivity are not sufficiently distinguished. Much of the analysis relies on hypothetical scenarios rather than nuanced engagement with the data.

 

While the author acknowledges that the lack of detailed analysis is a limitation of a comprehensive review, the current version remains too vague. In Section 4.1, the connection between descriptions and identified benefits needs clarification. At present, the relevance to inclusivity is unclear.

 

In Table 1 (Online learning and distance education), the links between descriptions and potential benefits are weak. How AI can enhance accessibility is not adequately explained, nor is it clear how AI mitigates geographical and temporal barriers based on the descriptions provided.

 

In addition, the study should demonstrate a sharper awareness of minority groups in education if inclusivity is to be its central focus. Who counts as minorities, and how can AI-supported education address their needs? Including specific cases or examples – jeven one or two – would strengthen the argument. For instance, how might adaptive learning and personalization benefit language minorities, students with disabilities, or sexual minorities?

 

In Section 4.4.6, the discussion of the emotional and affective dimension of AI does not fully reflect current concerns. Strong emotional attachments to AI and over-reliance on technology have raised new issues. In educational settings, some studies have shown that students may find interactions with AI more comfortable than with lecturers or peers. These nuances should be incorporated into the discussion.

Author Response

Comment 1: This article examines a timely and important topic – inclusivity in AI use in educational settings – and offers a comprehensive overview of AI applications in higher education. However, the manuscript requires extensive revision before it can be considered publishable.
Response 1: We thank you for your thorough review and constructive feedback. We will address all comments below.

Comment 2: The theoretical section needs to more clearly establish the research gap. At present, it is unclear how this study is conceptually and theoretically situated in relation to previous studies. The section reads more like a literature review than a theoretical framework, which is particularly confusing given that the methodology is a rapid review. This section would benefit from a revised title and clearer articulation of the research questions in relation to prior studies.
Response 2: To clarify the section's purpose within a rapid review framework, we have retitled it from "Theoretical Background and Practical Relevance" to "2. Situating the research: Background and relevance." The intention of this section is not to develop a novel theoretical framework but to reviewing existing literature to clearly contextualise and articulate the research question. The change can be found on line 81.

Comment 3: Furthermore, the analysis does not appear to emerge naturally from the review but instead feels predetermined and overly structured. One reason is that the general benefits of AI and the potential benefits for inclusivity are not sufficiently distinguished. Much of the analysis relies on hypothetical scenarios rather than nuanced engagement with the data.
Response 3: We have taken this important feedback to heart and made two key changes. First, we have restructured Table 1 in line 288 to include two distinct columns for general benefits and benefits for inclusivity. Second, we have added a new introductory paragraph to Section 4 (Results) in line 252 until line 262 that explains the analytical logic of our method and section structure, clarifying that it is a deliberate development of insights based on the data selected in the review.

Comment 4: While the author acknowledges that the lack of detailed analysis is a limitation of a comprehensive review, the current version remains too vague. In Section 4.1, the connection between descriptions and identified benefits needs clarification. At present, the relevance to inclusivity is unclear.
Response 4: We agree with this point and have addressed it alongside the previous feedback point on Table 1 that starts on line 288.

Comment 5: In Table 1 (Online learning and distance education), the links between descriptions and potential benefits are weak. How AI can enhance accessibility is not adequately explained, nor is it clear how AI mitigates geographical and temporal barriers based on the descriptions provided.
Response 5: The revised Table 1 on line 288 now features a dedicated column, "Potential benefits for inclusive education," with each entry explaining how a general capability can serve inclusive ends.

Comment 6: In addition, the study should demonstrate a sharper awareness of minority groups in education if inclusivity is to be its central focus. Who counts as minorities, and how can AI-supported education address their needs? Including specific cases or examples – even one or two – would strengthen the argument. For instance, how might adaptive learning and personalization benefit language minorities, students with disabilities, or sexual minorities?
Response 6: We agree. Our definition of "minorities" is established in Section 2, in the paragraph beginning on line 87. To reinforce this, the revised Table 1 on line 288 describes benefits in terms of principles that apply to these groups, and our Conclusion (Section 6) includes more concrete recommendations in the paragraph beginning on line 781 until line 788.

Comment 7: In Section 4.4.6, the discussion of the emotional and affective dimension of AI does not fully reflect current concerns. Strong emotional attachments to AI and over-reliance on technology have raised new issues. In educational settings, some studies have shown that students may find interactions with AI more comfortable than with lecturers or peers. These nuances should be incorporated into the discussion.
Response 7: We have revised Section 4.4.6 (Emotional and Social Impact) from line 575 to 580 to incorporate this nuanced perspective. We acknowledge that some students, particularly those with social anxiety, may find AI interactions more comfortable, while still weighing this against the broader concerns about the development of social skills and independence and the importance of human mentorship.

Reviewer 2 Report

Comments and Suggestions for Authors

This article examines the impact of AI applications on the inclusivity of minority groups in higher education using a rapid review approach. Synthesizing academic publications, policy reports, and case studies reveals that current tools focus on institutional efficiency, insufficiently address inclusivity, and involve significant ethical challenges.

Strengths

• Timely and significant topic: Addresses the impact of AI on inclusivity in higher education.
• Broad data sources: Offers multiple perspectives by incorporating academic articles, policy reports, and case studies.
• Structured thematic analysis: Employs inductive coding organized around four main themes, reinforcing the logical progression of findings.
• Transparency statement: Clearly discloses the use of generative AI tools and emphasizes authors’ accountability, demonstrating a strong ethical stance.

However, I have major comments as below:

• Limited scope: The rapid review evaluates only the first ten results per search, which falls short of a full systematic review standard.
• Lack of methodological quality assessment: No risk-of-bias or quality scoring of included studies is provided, potentially undermining the review’s reliability.
• Absence of primary data: Findings rely entirely on secondary sources without incorporating direct perspectives from students, faculty, or administrators.
• Limited theoretical and operational innovation: The study remains at the level of high-level recommendations rather than developing a concrete theoretical model or measurement tool.

Moreover, I have minor comments as below:

• Update placeholder information for “Received, Revised, Accepted, Published” dates.
• Ensure consistency in numbering of section headings (e.g., between 3.3.2 and 3.3.3).
• Clarify definitions of terms such as “rapid review” and “thematic analysis” once more within the text.
• Specify inclusion/exclusion criteria for case studies and policy reports more explicitly.

Author Response

Comment 1: This article examines the impact of AI applications on the inclusivity of minority groups in higher education using a rapid review approach. Synthesizing academic publications, policy reports, and case studies reveals that current tools focus on institutional efficiency, insufficiently address inclusivity, and involve significant ethical challenges.

Strengths

• Timely and significant topic: Addresses the impact of AI on inclusivity in higher education.
• Broad data sources: Offers multiple perspectives by incorporating academic articles, policy reports, and case studies.
• Structured thematic analysis: Employs inductive coding organized around four main themes, reinforcing the logical progression of findings.
• Transparency statement: Clearly discloses the use of generative AI tools and emphasizes authors’ accountability, demonstrating a strong ethical stance.

Response 1: We thank the reviewer for the thorough and constructive feedback. We have actioned all the items and provide further information on the points below.

Comment 2: Limited scope: The rapid review evaluates only the first ten results per search, which falls short of a full systematic review standard.
Response 2: We thank the reviewer for this critical point. We have revised Section 3.3 to clarify that our search was a complementary process: an initial scan of top results was augmented with citation network analysis to ensure influential works were included.  A paragraph was added on line 200.

Comment 3: Lack of methodological quality assessment: No risk-of-bias or quality scoring of included studies is provided, potentially undermining the review’s reliability.
Response 3: The reviewer is correct. We have explicitly addressed this as a deliberate feature of the rapid review methodology in our Limitations Section (6.1) starting on line 801.

Comment 4: Absence of primary data: Findings rely entirely on secondary sources without incorporating direct perspectives from students, faculty, or administrators.
Response 4: We have created a dedicated subsection in our revised limitations section starting on line 805 to discuss this point. This section now clearly states that the absence of firsthand accounts is a key limitation and is further reinforced by a stronger call for primary research in the Future Work section (6.2) from line 835 to line 842.

Comment 5: Limited theoretical and operational innovation: The study remains at the level of high-level recommendations rather than developing a concrete theoretical model or measurement tool.
Response 5: To make our recommendations less high-level, we made two key additions: a "practical heuristic" for institutions in the Discussion (Section 5.1) from line 672 to line 679 and a list of concrete, actionable steps in the Conclusion (Section 6) from line 781 to line 788.

Comment 6: Update placeholder information for “Received, Revised, Accepted, Published” dates.
Response 6: Thank you. This information will be added by the editorial staff during the production phase, as noted in the manuscript.

Comment 7: Ensure consistency in numbering of section headings (e.g., between 3.3.2 and 3.3.3).
Response 7: We have thoroughly reviewed all section headings and subheadings for numbering and formatting consistency and have corrected any errors.

Comment 8: Clarify definitions of terms such as “rapid review” and “thematic analysis” once more within the text.
Response 8: We have expanded the definition of "rapid review" in Section 3.2 from line 156 to line 164 and included a formal citation to provide a stronger methodological grounding.

Comment 9: Specify inclusion/exclusion criteria for case studies and policy reports more explicitly.
Response 9: We have revised the methodology section to address this. The core inclusion criteria are now explicitly stated in Section 3.3. from line 192 to line 204.

Reviewer 3 Report

Comments and Suggestions for Authors

As the authors themselves acknowledge, the exclusive reliance on secondary data constitutes a notable limitation, as it restricts the depth of analysis regarding how AI tools are experienced by key stakeholders. The absence of first-hand perspectives from students, instructors, and administrative staff may hinder the study’s ability to fully capture the practical implications of AI integration within educational contexts. While this limitation does not diminish the overall value of the paper, it represents an important area for improvement in future research.

Nevertheless, the study offers a systematic and comprehensive examination of a pressing contemporary issue. By synthesizing academic literature, policy reports, and case studies, it presents a holistic perspective and effectively integrates ethical, pedagogical, and social dimensions of AI use. The conceptual framework is well-constructed, and the incorporation of diverse source types enhances the academic rigor of the work. That said, the conclusions would benefit from being supported by more concrete, context-specific, and actionable recommendations. Strengthening the practical implications of the findings would enhance both the theoretical contribution and the relevance of the study for policymakers and practitioners.

Additionally, the manuscript predominantly draws upon literature from Western contexts. While this is explicitly acknowledged as a limitation, the absence of examples from the Global South restricts the generalizability of the findings. Integrating perspectives from underrepresented educational systems could offer valuable insights into challenges such as digital inequality, data representation, and algorithmic bias in diverse sociocultural settings.

In sum, despite its methodological limitations, the paper makes a valuable contribution to the field and brings attention to critical discussions surrounding the future of higher education in the age of AI. With minor revisions particularly regarding the articulation of practical recommendations the manuscript would be suitable for publication.

Author Response

Comment 1: As the authors themselves acknowledge, the exclusive reliance on secondary data constitutes a notable limitation, as it restricts the depth of analysis regarding how AI tools are experienced by key stakeholders. The absence of first-hand perspectives from students, instructors, and administrative staff may hinder the study’s ability to fully capture the practical implications of AI integration within educational contexts. While this limitation does not diminish the overall value of the paper, it represents an important area for improvement in future research.
Response 1: We thank the reviewer and fully agree. We have dedicated a paragraph in the Limitations section (6.1) to discussing the limitations of relying on secondary data and the lack of stakeholder perspectives. This can be found from line 807 until line 833.

Comment 2: Nevertheless, the study offers a systematic and comprehensive examination of a pressing contemporary issue. By synthesizing academic literature, policy reports, and case studies, it presents a holistic perspective and effectively integrates ethical, pedagogical, and social dimensions of AI use. The conceptual framework is well-constructed, and the incorporation of diverse source types enhances the academic rigor of the work. That said, the conclusions would benefit from being supported by more concrete, context-specific, and actionable recommendations. Strengthening the practical implications of the findings would enhance both the theoretical contribution and the relevance of the study for policymakers and practitioners.
Response 2: We are grateful for this feedback. We have revised the Conclusion (Section 6) to include a paragraph with specific, actionable recommendations, such as the formation of "AI ethics committees". This can be found from line 781 to line 788.

Comment 3: Additionally, the manuscript predominantly draws upon literature from Western contexts. While this is explicitly acknowledged as a limitation, the absence of examples from the Global South restricts the generalizability of the findings. Integrating perspectives from underrepresented educational systems could offer valuable insights into challenges such as digital inequality, data representation, and algorithmic bias in diverse sociocultural settings.
Response 3: 
We have elevated this point to a core part of our revised Limitations Section (6.1) starting on line 803 until line 833, where we now explicitly discuss the geographic and English-language bias and its impact on generalisability.

Comment 4: In sum, despite its methodological limitations, the paper makes a valuable contribution to the field and brings attention to critical discussions surrounding the future of higher education in the age of AI. With minor revisions particularly regarding the articulation of practical recommendations the manuscript would be suitable for publication.
Response 4: We sincerely thank the reviewer for their positive assessment and valuable guidance. We believe the extensive revisions made in response to all reviewers' comments have substantially strengthened the manuscript and addressed the highlighted concerns in depth.

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

In line with my previous comments, I find the progress of the work and the authors' responses satisfactory and sufficient for publication.