Exploring Student Acceptance of AI Teaching Assistants in African Higher Education

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Thank you for submitting this manuscript to Trends in Higher Education. This study addresses a timely and underexplored topic: student acceptance of AI teaching assistants within a South African clinical learning environment. Your use of the Technology Acceptance Model (TAM) as a theoretical lens provides a purposeful and coherent framework for the inquiry. The evaluation below assesses the manuscript across six key criteria, identifying both the strengths your work already demonstrates and the specific areas where revision would meaningfully strengthen the manuscript before publication.

The manuscript's most significant strength is its contextual specificity. AI adoption research in education is dominated by studies from North America, Europe, and East Asia. By situating this study in a South African clinical learning environment, you make a genuine contribution to an underrepresented area of the literature. This geographic and disciplinary positioning is the manuscript's clearest point of originality and should be foregrounded more explicitly throughout the paper, particularly in the introduction and conclusion.

The qualitative single-case study design within an interpretivist paradigm is well-suited to the research's exploratory aims. Your methodological rationale is clearly articulated, and the purposive sampling strategy is appropriate for the study context. The inclusion of trustworthiness strategies, member checking, and an audit trail demonstrates sound methodological practice and strengthens the credibility and confirmability of your findings. These elements reflect a solid command of qualitative research design.

The limitations section is commendably transparent. You acknowledge clearly that the small sample size, single institutional context, and self-reported data are constraints on the study's transferability and depth. This honesty is an important scholarly virtue and ensures that readers engage with your findings appropriately. The inclusion of a practical recommendations section also adds applied value for educators and institutional decision-makers.

However, the following areas require substantive attention in the revised manuscript. Each concern is paired with a specific recommendation.

1-

The TAM is applied in a largely conventional manner, without meaningful contextual adaptation. While TAM is a well-established and appropriate theoretical lens, its standard application here does not produce sufficient new theoretical insight for the African higher education context. The distinctive challenges of this setting, including unequal digital infrastructure, varying levels of digital literacy, limited device access, and the specific demands of clinical education , represent variables that are known to shape technology acceptance but are not integrated into the theoretical model.

Recommendation: Extend the TAM framework by incorporating additional constructs relevant to the South African or broader African context. This might include infrastructure readiness, institutional trust, or digital equity as antecedents or moderators of perceived usefulness and ease of use. Alternatively, if you choose to retain a standard TAM application, provide a more compelling and explicit theoretical argument for why the standard model is analytically sufficient in this context.

2-

While the manuscript makes a compelling case for the importance of context-specific research on AI adoption in African higher education, the literature review draws primarily from global AI-in-education scholarship. Claims about digital inequality, infrastructural barriers, and technology adoption in South African universities are supported with general references, but a more targeted engagement with African-specific literature is needed to appropriately situate your study within the regional scholarly conversation.

Recommendation: Expand engagement with African-specific scholarship across three areas relevant to your study: digital inequality and infrastructure disparities in South African or sub-Saharan African universities; technology adoption and student attitudes in African higher education contexts; and clinical education and professional training in resource-limited or developing country settings. This will both strengthen the theoretical grounding of the study and demonstrate deeper familiarity with the regional research landscape.

3-

The discussion currently summarizes the findings and confirms their alignment with prior literature. While this provides useful contextualization, it does not fulfill the full analytical purpose of a discussion section. Readers expect the discussion to move beyond summary toward critical interpretation, identifying unexpected findings, surfacing tensions between data and theory, and articulating implications that challenge or extend current knowledge.

Recommendation: Restructure the discussion to foreground at least two or three interpretive claims that go beyond confirming prior findings. Consider questions such as: What does it mean that students in a resource-constrained African clinical setting express largely positive attitudes toward AI despite infrastructure limitations? How does the clinical education context create specific demands on AI tools that general higher education settings do not? What are the implications for how AI tools should be designed, deployed, and supported in similar settings?

4-

The abstract contains a grammatical error in its opening sentence: 'Artificial intelligence (AI) is increasingly being integrated into higher education Among these innovations...' A period or transitional phrase is missing between the two clauses. Please correct this before resubmission.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

Thank you for the opportunity to revise and resubmit the manuscript. I sincerely appreciate the constructive feedback provided by the reviewer, which has significantly strengthened the quality and clarity of the paper.

In response to the comments, I have carefully revised the manuscript. Specifically, I have clarified the justification for the application of the Technology Acceptance Model (TAM) within the African higher education context, incorporated relevant African-specific scholarship to strengthen the regional grounding of the study, and restructured the discussion section to provide deeper analytical interpretation and highlight key contributions. Additionally, the manuscript has undergone professional language editing, and all identified grammatical issues have been corrected.

I believe that these revisions have substantially improved the manuscript and addressed the reviewer’s concerns. A detailed, point-by-point response is provided in the table.

Thank you for your consideration. I look forward to your feedback.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I agree with the authors that AI teaching assistants are becoming increasingly important in higher education, their role will only grow. The aim of the study is to investigate the knowledge, attitudes, and acceptance of AI teaching assistants in one university. The theoretical basis in the study has been successfully selected and presented, with an attempt to link it to the results obtained.

However, there are three main objections, two of which are interrelated.

• Inappropriate references used in the literature review. Of course, I didn't read all the sources mentioned in the manuscript, but I found that: - In the 2nd source, which is referenced after the sentence in lines 102-104, there is no mention of artificial intelligence; - source 21 is about artificial intelligence in the healthcare sector, there is nothing about higher education and AI assistants (line 112-113); source 24 is about “Artificial Intelligence in Tongue Image Recognition”, not about academic support (line 121-122). Therefore, there is no credibility to the literature review.
• In the manuscript, the study design is referred to a single case study. There is a reference to source 48 (Yin’s textbook) in Part 4 Methods. Unfortunately, the author has ignored an essential parameter of the case study, which Yin describes on page 46: “[Case study] relies on multiple sources of evidence, with data needing to converge in a triangulating fashion”. The data for this study are only 6 semi-structured interviews; there is only one data source. It is inappropriate to call this study a case study, it should at least include the lecturers' perspective. Additionally, the explanation for the small number of interviews (lines 283-284) is rather naive and unconvincing.
• The section Method for case study design also refers to source 13, which I managed to find and read. Subchapter in this manuscript 4.2. Sampling Technique and Participants is very similar to the 13th source subsection "Study participants and sampling ". Plagiarism is suspected because the sentences have been minimally reworded. Yes, research tends to replicate methods and designs, but then it is necessary to appropriately cite that the design was copied from such a source.

In general, the journal's requirements have been met, except that I would recommend including references in sentences, as is done in the journal's publications. Overall, this is a very minor study, it could be like an attempt at a first publication. Then it is understandable to try to copy a design described in another publication by casually reading a textbook on case studies.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

This study addresses a timely and relevant topic in relation to ongoing transformations in higher education and has the merit of drawing attention to students’ perceptions of AI teaching assistants in an underexplored context, namely clinical education in South African higher education. However, the manuscript would benefit from several important revisions.

The abstract is well structured, but it remains somewhat generic in tone. It would benefit from a clearer articulation of the study’s specific contribution, a stronger alignment between TAM and the summarized findings, and a more cautious presentation of the conclusions given the limited scope for broader generalization.

The introduction, literature review, and theoretical framework provide a coherent conceptual foundation and a logical progression from digital transformation in African higher education to AI teaching assistants, AI literacy, and technology acceptance. However, this section remains more descriptive than analytical. The research gap is formulated too broadly, the study’s contribution is not sufficiently distinguished from related work, and the broader African framing is not fully aligned with the actual single-case South African clinical context. The integration of TAM would also benefit from a clearer alignment with the research question and the qualitative focus of the study.

The methodology section is the weakest part of the manuscript, as it does not provide enough detail to demonstrate methodological rigor. The AI teaching assistant intervention is not described with sufficient precision, the rationale for case selection is not entirely convincing, and the inclusion of only six participants is not adequately discussed in relation to qualitative adequacy or thematic saturation. In addition, the procedures for data collection and thematic analysis are reported too briefly, making it difficult to assess how the interviews were conducted, how codes and themes were generated, and how interpretive consistency was ensured. The section would also benefit from a clearer account of researcher reflexivity and of how credibility-enhancing strategies, such as member checking and maintaining an audit trail, were implemented.

The results section is clearly structured and broadly aligned with the aim of the study. However, the analysis remains largely descriptive, with some overlap between themes, limited empirical depth, and only modest exploration of variation or divergent perspectives across participants. The clinical context is mentioned but not sufficiently developed as a distinct analytical feature, and the alignment with TAM remains weak. Moreover, although the authors state that they conducted a thematic analysis, the analytical structure of the findings is not made sufficiently explicit. As presented, the findings read more like a thematic summary than a fully demonstrated thematic analysis.

The discussion section is coherent and relevant, but it tends to restate the results rather than deepen them interpretively. Its engagement with the literature remains largely confirmatory, the study’s specific contribution is not sufficiently problematized, and the alignment with TAM is only partial. In addition, the clinical and African/South African contexts are mentioned but not developed enough as analytically distinctive dimensions. Although the manuscript acknowledges limitations in a separate section, the discussion would be strengthened by integrating these limitations more directly into the interpretation of the findings and by adopting a more cautious tone regarding generalization.

The conclusion and recommendations are relevant and consistent with the overall direction of the study. However, the conclusion remains more recapitulatory than synthetic-interpretive and does not highlight the study’s distinct contribution with sufficient clarity. The recommendations are generally pertinent, especially those related to AI literacy, the complementary use of AI, and institutional guidance, but they are expressed in rather broad and normative terms and at times extend beyond the study’s direct empirical base, given that only students’ perspectives were examined.

Best regards,

Comments on the Quality of English Language

 The English could be improved to more clearly express the research.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

This study makes a meaningful contribution to the underexplored area of AI adoption in African higher education, particularly by grounding its inquiry in a clinical learning context where few such investigations exist. The use of the Technology Acceptance Model as a theoretical lens is well-justified, and the thematic analysis is clearly structured around the three emergent themes. To strengthen the work further, consider expanding the discussion of how the interpretivist paradigm specifically shaped your interpretations, and reflect more explicitly on the researcher's positionality, given your direct access to the learning environment and participants. The sample size of six participants, while consistent with qualitative traditions, warrants a more robust justification in the methods section, and the limitation noted regarding the absence of educator perspectives could be partially addressed by discussing what insights were consequently missed. A minor typographical error ("qualitative," p. 6) should also be corrected before the submission.

Author Response

The interpretivist paradigm shaped the analysis by prioritising participants’ sub-jective meanings and lived experiences of engaging with AI teaching assistants within their clinical learning context. Rather than seeking objective measurement of technology acceptance, the study focused on how students interpreted the usefulness, limitations, and role of AI in relation to their own learning experiences. The researcher’s positionality also influenced the study, as direct access to the learning environment facilitated deeper contextual understanding and rapport with participants. At the same time, reflexive awareness was maintained to minimise potential bias, particularly during data inter-pretation, by using member checking, maintaining an audit trail, and critically reflecting on how prior assumptions may have shaped the analysis.

 

 

 

The following discussion was added (see the revised manuscript, highlighted)

 

The sample size of six participants was considered appropriate for this qualitative study because the aim was to generate in-depth, context-specific insights rather than broad generalisations. Data adequacy was determined by the richness, relevance, and repetition of participant responses in relation to the research question, rather than by numerical representation alone.

 

 

Although the findings provide valuable student perspectives, the absence of educator and clinical supervisor voices means that important insights regarding teaching practices, professional judgment, and institutional readiness for AI integration may not have been fully captured. Including these perspectives could have provided a more comprehensive understanding of how AI teaching assistants are positioned within clinical education and how both students and educators negotiate their use in practice. Future studies could include multiple stakeholders and larger samples to provide a more comprehensive understanding of AI adoption in higher education.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have worked on text improvements, and most of the suggestions have been considered. The exception is Chapter 4, Methods, where lines 274-282 still describe the case. Here is again my comment from the previous review: “There is a reference to source 48 (Yin’s textbook) in Part 4 Methods. Unfortunately, the author has ignored an essential parameter of the case study, which Yin describes on page 46: “[Case study] relies on multiple sources of evidence, with data needing to converge in a triangulating fashion”. The data for this study are only 6 semi-structured interviews; there is only one data source. It is inappropriate to call this study a case study, it should at least include the lecturers' perspective.”

Author Response

My apologies for overlooking this section.

I fully agree that the use of the term “case study” was not methodologically appropriate in this context. The entire case study description, including the reference to Yin’s case selection criteria, has been removed to ensure better alignment with the descriptive qualitative design and interpretivist paradigm adopted in this study. Please see Lines 269–278 for the revised section.

 

To further strengthen the manuscript, additional discussion has also been included in the limitations section regarding the absence of other stakeholders’ perspectives, particularly lecturers and clinical supervisors, and the insights that may have been missed as a result (see Section 9 [highlighted]).

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I appreciate the authors’ efforts to substantially revise the manuscript by clarifying the study’s contribution, more rigorously delimiting the South African clinical context, strengthening the theoretical framework, and introducing a more explicit interpretation of the results and discussion. These changes indicate a genuine revision and a serious attempt to address the concerns raised previously.

That said, I believe the aspect that still requires the greatest attention is the rigor and transparency of the thematic analysis, as this constitutes the foundation of credibility in a qualitative study. Although the authors state that they used a six-step thematic analysis and refer to strategies such as member checking and an audit trail, these elements are still presented at a general rather than a procedural level. The manuscript does not explain clearly enough how the initial codes were generated, how the themes were developed and refined, whether the coding process was inductive, deductive, or partially guided by the TAM framework, who carried out the coding, and how interpretive consistency was ensured throughout the analysis. In its current form, the reader can identify the final themes, but cannot follow with sufficient transparency the analytical pathway through which these themes were constructed from the raw data. In addition, researcher reflexivity remains insufficiently discussed, even though the researchers’ proximity to the research context would justify a clearer reflection on the potential influence of the researcher on both data collection and interpretation. Likewise, the justification for the six-participant sample, although better articulated than in the previous version, is still not demonstrated with sufficient methodological clarity.

The results are more interpretive than in the previous version, which represents a real improvement. However, a rigorous thematic analysis requires not only more analytical wording, but also a clearer and more transparent link between data, codes, themes, and interpretation. In the current version, this traceability remains only partially visible, and variations across participants, as well as any potentially divergent perspectives, are not explored sufficiently. 

Comments on the Quality of English Language

 The English could be improved to more clearly express the research.

Author Response

Thank you for this valuable comment.

Section 4.4 (Data Analysis) has been substantially revised to improve methodological transparency and provide a clearer procedural description of the analytical process. The manuscript now explains how initial codes were generated through line-by-line analysis, how themes were developed and refined using Braun and Clarke’s six-step thematic analysis, and how the coding process followed a hybrid deductive–inductive approach, with the Technology Acceptance Model serving as a sensitising framework while allowing new themes to emerge from the data.

Additional clarification has also been provided regarding who conducted the coding, how interpretive consistency was strengthened through peer debriefing and repeated comparison with the original transcripts, and how researcher reflexivity was maintained through reflexive journaling, member checking, and an audit trail. These revisions improve the transparency and trustworthiness of the analytical pathway from raw data to final themes.

See Section 4.4, Line 308-333.

 

Thank you for this valuable comment. The Results section has been further revised to improve the transparency and traceability of the thematic analysis. Additional explanation has been included to clarify how initial codes were generated, grouped into categories, and developed into the final themes and subthemes. To strengthen methodological transparency and provide a clearer link between raw data, coding, and interpretation, a full coding summary table has also been added as an Appendix. This table presents the relationship between initial codes, categories, themes, illustrative participant quotations, and the corresponding interpretations, allowing readers to more clearly follow the analytical pathway from interview data to final thematic findings.

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

The revised manuscript has substantially addressed the main methodological concerns raised in the previous review, particularly regarding the transparency and rigor of the thematic analysis. The authors now provide a clearer account of the coding process, the deductive–inductive analytical approach, the role of TAM as a sensitising framework, peer debriefing, member checking, reflexivity, and the audit trail. The addition of a coding summary appendix also strengthens the traceability between raw data, codes, themes, quotations, and interpretation.  However, the manuscript requires careful final proofreading and editorial cleaning, as several track-change remnants, duplicated words, formatting markers, citation-format errors, and a possible date inconsistency remain in the text.

Comments on the Quality of English Language

 The English could be improved to more clearly express the research.