Review Reports - Artificial Intelligence Adoption and the Need for Artificial Intelligence Literacy in Veterinary Education: A Cross-Sectional Survey

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Overall, the topic is relevant, but the manuscript requires substantial revision, particularly with respect to its positioning within previous research, methodological transparency, the development of the Discussion section.

The manuscript lacks a dedicated Related Work section. Given the substantial and rapidly growing body of literature on AI adoption in education, the authors should position their study more clearly within the existing research landscape. Relevant findings from previous studies should be critically discussed rather than merely cited. The manuscript also refers to several studies concerning AI use in veterinary education. These studies should be discussed in greater depth to establish the specific context of the present research, identify what remains insufficiently investigated, and clearly demonstrate the scientific gap and novelty of the study. More generally, the manuscript would benefit from a stronger discussion of the current state of AI adoption specifically in veterinary education. This context is necessary to explain why the study is important for this particular educational domain.
The description of the methodology requires substantial improvement. First, the authors should clearly formulate the aim of the study and provide explicit research questions. Second, the manuscript reports the sample size but does not clearly specify the size and characteristics of the target population. This information is necessary for evaluating the representativeness of the sample and interpreting the reported prevalence estimates. The development of the questionnaire should also be described in greater detail. At present, the manuscript mainly provides references to sources that inspired questionnaire development. The authors should explain how the questionnaire was constructed, how individual items were selected or adapted, whether experts were involved in reviewing the instrument, and whether any pilot testing or validation procedures were performed.
Although the manuscript contains a section entitled Discussion, it does not currently function as a sufficiently developed scientific discussion. The results are primarily restated rather than systematically compared and contrasted with findings from previous studies. This issue is closely related to the absence of a well-developed Related Work section. A meaningful discussion should identify where the present findings agree with previous research, where they differ, and what possible explanations exist for these similarities or differences. It should also explain the implications of the findings for veterinary education and AI adoption. One indication of the current problem is the frequent placement of references only at the ends of paragraphs. While citation placement is not itself a methodological issue, a more analytical discussion would normally integrate previous studies directly into the argument-for example, by explicitly comparing particular findings of the present study with those reported by other researchers.
The study design allows the authors to describe prevalences and identify associations. However, it does not support conclusions regarding causal relationships. The manuscript should therefore be carefully revised to remove or rephrase any terminology suggesting cause-and-effect relationships.
Minor drawbacks:

Good academic practice generally recommends avoiding abbreviations in the title of a paper. Consider spelling out an abbreviated term in the title.
Abbreviations should be used consistently throughout the manuscript. A term should normally be written in full at its first occurrence, followed by the abbreviation in parentheses; thereafter, only the abbreviation should be used. In the present manuscript, for example, both “artificial intelligence” and “AI” appear inconsistently after the abbreviation has already been introduced.
Consider avoiding the combination of numerous plots or figures into a single figure with very long figure captions. Where appropriate, separate figures may improve readability.
Consider replacing first-person expressions such as “we” and “our” with more neutral academic formulations where appropriate.
Tables 5, 6, and 7 are referenced in the text, but I could not locate them in the manuscript. Please check that all tables are included and correctly numbered.
There is generally no need to provide the full legal or corporate names of companies that developed particular software products.
At the end of the Introduction, add a brief paragraph describing the structure of the remainder of the paper.

Author Response

We sincerely thank the reviewer for the careful evaluation of our manuscript and for the constructive and insightful comments. These suggestions have been invaluable in improving the quality, clarity, and scientific rigor of the work. We have thoroughly considered all comments and have revised the manuscript accordingly. Below, we provide a detailed, point-by-point response to each observation, indicating the changes made and the corresponding sections of the revised manuscript where appropriate.

Comment 1: The manuscript lacks a dedicated Related Work section. Given the substantial and rapidly growing body of literature on AI adoption in education, the authors should position their study more clearly within the existing research landscape. Relevant findings from previous studies should be critically discussed rather than merely cited. The manuscript also refers to several studies concerning AI use in veterinary education. These studies should be discussed in greater depth to establish the specific context of the present research, identify what remains insufficiently investigated, and clearly demonstrate the scientific gap and novelty of the study. More generally, the manuscript would benefit from a stronger discussion of the current state of AI adoption specifically in veterinary education. This context is necessary to explain why the study is important for this particular educational domain.

Response: We thank the reviewer for this valuable comment. We agree that the original manuscript did not sufficiently position the present study within the rapidly expanding literature on AI adoption in higher education, particularly within veterinary education. Although previous studies were cited in the Introduction, their findings were not discussed in sufficient depth to clearly establish the specific research gap addressed by the present study. We have therefore substantially revised the Introduction to provide a more focused and critical synthesis of previous research on AI use in higher education and, more specifically, in veterinary education. The revised text now discusses reported patterns of AI adoption, students’ knowledge and attitudes, perceived educational benefits, concerns regarding reliability and academic integrity, and the limited availability of formal AI training and institutional guidance. We have also expanded the discussion of previous studies involving veterinary students to clarify what has already been investigated and which aspects remain insufficiently characterized. In particular, we emphasize the limited evidence regarding the combined assessment of AI knowledge, usage patterns, perceptions, educational impact, ethical considerations, institutional guidance, and future expectations among veterinary students within a single study context. This revised synthesis allows the rationale and contribution of the present study to be more clearly established. We have incorporated these changes into the Introduction rather than creating a separate Related Work section, as we considered that integrating the literature review with the study rationale provides a more coherent structure for this type of empirical educational study.

Comment 2: The description of the methodology requires substantial improvement. First, the authors should clearly formulate the aim of the study and provide explicit research questions. Second, the manuscript reports the sample size but does not clearly specify the size and characteristics of the target population. This information is necessary for evaluating the representativeness of the sample and interpreting the reported prevalence estimates. The development of the questionnaire should also be described in greater detail. At present, the manuscript mainly provides references to sources that inspired questionnaire development. The authors should explain how the questionnaire was constructed, how individual items were selected or adapted, whether experts were involved in reviewing the instrument, and whether any pilot testing or validation procedures were performed.

Response: We thank the reviewer for this important observation. We agree that the original description did not provide sufficient detail regarding the development of the questionnaire. The questionnaire was developed de novo specifically for this study rather than adapted from a previously validated instrument. Its content was informed by a review of recent literature on AI use in higher education, digital competence, and AI applications in medical and veterinary education. The literature was used to identify the main domains relevant to the study objectives and to guide the formulation of the individual items. The resulting questionnaire was subsequently reviewed by the research team to assess whether the items adequately covered the relevant aspects of AI use and its educational implications in veterinary education. We have expanded Section 2.4 to clarify this development process and the rationale underlying the structure and content of the questionnaire. No formal pilot testing or psychometric validation was performed, as the instrument was developed as a study-specific descriptive questionnaire rather than as a validated scale designed to measure a single latent construct.

Comment 3: Although the manuscript contains a section entitled Discussion, it does not currently function as a sufficiently developed scientific discussion. The results are primarily restated rather than systematically compared and contrasted with findings from previous studies. This issue is closely related to the absence of a well-developed Related Work section. A meaningful discussion should identify where the present findings agree with previous research, where they differ, and what possible explanations exist for these similarities or differences. It should also explain the implications of the findings for veterinary education and AI adoption. One indication of the current problem is the frequent placement of references only at the ends of paragraphs. While citation placement is not itself a methodological issue, a more analytical discussion would normally integrate previous studies directly into the argument-for example, by explicitly comparing particular findings of the present study with those reported by other researchers.

Response: We thank the reviewer for this important and constructive comment. We agree that the original Discussion relied too heavily on a description of the main findings and that the comparison with previous studies was not sufficiently systematic. In response, we have substantially revised the Discussion to provide a more analytical interpretation of the findings. For the main results, we now explicitly compare our findings with previous studies, highlighting areas of agreement and differences and, where appropriate, discussing possible explanations for these patterns. We have also strengthened the discussion of the implications of our findings for veterinary education, particularly regarding AI literacy, responsible AI use, critical evaluation of AI-generated information, institutional guidance, and the integration of AI into veterinary curricula. In addition, references have been incorporated more directly into the relevant arguments rather than being placed only at the end of paragraphs. These revisions aim to distinguish more clearly between findings that are consistent with previous research, findings that may reflect the specific educational context of our institution, and interpretations that should be considered cautiously given the cross-sectional and self-reported nature of the study.

Comment 4: The study design allows the authors to describe prevalences and identify associations. However, it does not support conclusions regarding causal relationships. The manuscript should therefore be carefully revised to remove or rephrase any terminology suggesting cause-and-effect relationships.

Response: We thank the reviewer for this important methodological observation. We agree that the cross-sectional design of the present study allows us to describe AI use, perceptions, and attitudes and to examine associations between variables, but does not permit causal inference. We have therefore carefully reviewed the manuscript and revised the terminology where necessary to avoid implying cause-and-effect relationships. In particular, statements referring to AI as “improving,” “influencing,” or “leading to” particular educational outcomes have been rephrased where appropriate to reflect students’ perceptions and reported experiences rather than objectively demonstrated effects. We have also clarified in the Limitations section that the cross-sectional design precludes causal conclusions between AI use, perceived learning benefits, and educational outcomes.

Minor drawbacks

Comment 1: Good academic practice generally recommends avoiding abbreviations in the title of a paper. Consider spelling out an abbreviated term in the title.

Response: Thank you for this helpful suggestion. We agree that avoiding abbreviations in the title improves clarity and accessibility for readers. Accordingly, the abbreviated term has been replaced with its full expression in the revised title.

Comment 2: Abbreviations should be used consistently throughout the manuscript. A term should normally be written in full at its first occurrence, followed by the abbreviation in parentheses; thereafter, only the abbreviation should be used. In the present manuscript, for example, both “artificial intelligence” and “AI” appear inconsistently after the abbreviation has already been introduced.

Response: Thank you for this helpful observation. We have carefully reviewed the entire manuscript to ensure the consistent use of abbreviations. Following your recommendation, all terms are now written in full at their first occurrence, followed by the corresponding abbreviation in parentheses. Subsequently, only the abbreviation is used throughout the manuscript. In particular, occurrences where both “artificial intelligence” and “AI” were used interchangeably after the abbreviation had already been introduced have been corrected to maintain consistency and improve readability.

Comment 3: Consider avoiding the combination of numerous plots or figures into a single figure with very long figure captions. Where appropriate, separate figures may improve readability.

Response: Thank you for this helpful suggestion. We have revised the presentation of the figures to improve readability and reduce the reliance on large composite figures with extensive captions. Where appropriate, figures have been separated into individual panels and the accompanying captions have been shortened and clarified. These modifications enhance the visual presentation of the results and facilitate interpretation of the data.

Comment 4: Consider replacing first-person expressions such as “we” and “our” with more neutral academic formulations where appropriate.

Response: As you recommend, we have carefully reviewed the manuscript and revised the language where appropriate to adopt a more neutral academic style. First-person expressions such as “we” and “our” have been replaced with impersonal or passive constructions when this improved the formality and consistency of the text. These modifications enhance the scholarly tone of the manuscript while maintaining clarity and readability.

Comment 5: Tables 5, 6, and 7 are referenced in the text, but I could not locate them in the manuscript. Please check that all tables are included and correctly numbered.

Response: Thank you for pointing this out. We carefully reviewed the manuscript and identified that the references to Tables 5, 6, and 7 were the result of an inconsistency in the numbering introduced during manuscript preparation. This issue has now been thoroughly checked and corrected throughout the manuscript. All figures are included in the revised version and are now consistently numbered and properly referenced in the text. We appreciate the reviewer’s attention to this issue, which has helped improve the clarity and accuracy of the manuscript.

Comment 6: There is generally no need to provide the full legal or corporate names of companies that developed particular software products.

Response: Thank you for this helpful suggestion. We have revised the manuscript accordingly by simplifying the software descriptions and removing unnecessary references to the full legal or corporate names of the companies associated with the software products. The software tools are now identified using their commonly recognized names, which improves readability while retaining all information necessary for reproducibility and transparency.

Comment 7: At the end of the Introduction, add a brief paragraph describing the structure of the remainder of the paper.

Response: Thank you for this suggestion. A brief paragraph describing the structure of the manuscript has been added at the end of the Introduction.

Reviewer 2 Report

Comments and Suggestions for Authors

This cross-sectional survey study investigated AI awareness, usage patterns, attitudes, and educational expectations among 189 undergraduate veterinary students at the University of Las Palmas de Gran Canaria, Spain. Key findings revealed near-universal AI adoption among students, predominantly ChatGPT, primarily for information retrieval, exam preparation, and academic support. Students expressed strong support for AI integration into the veterinary curriculum while simultaneously voicing substantial concerns about misinformation, overreliance, and impacts on critical thinking. The authors conclude that a significant gap exists between rapid AI adoption and formal AI literacy development in veterinary education, recommending curriculum integration of AI competencies, clearer institutional policies, and development of critical evaluation skills alongside technical proficiency. My observations and comments are as follows:

The non-uniform distribution Year 1: 28.6%, Year 2: 12.7%, Year 3: 6.3%, Year 4: 27.5%, Year 5: 24.9% is striking, particularly the underrepresentation of third-year students 6.3%. This raises questions about the representativeness of student perspectives across the curriculum and whether convenience sampling inadvertently excluded certain student populations, e.g., those on clinical rotations. The authors conduct chi-square tests showing significant differences (p<0.001) but do not adequately address how this uneven distribution might affect interpretation of year-specific patterns.
All data derive from self-report rather than objective assessments of AI literacy, academic performance, clinical reasoning, or critical thinking skills. The authors appropriately acknowledge this limitation, but the discussion frequently treats perceived benefits as if they were demonstrated outcomes. The statement that students reported improvements in conceptual understanding conflates subjective perception with objective learning gain.
The authors reference significant differences in demographic distributions but do not conduct association tests between demographics and AI-related outcomes.
The small sample size for AI non-users n=3 precludes meaningful subgroup analysis, yet the authors present percentages (66.7%, 33.3%) that could be misleading due to the tiny denominator.
Given the convenience sampling strategy and potential self-selection bias, how confident can we be that the 43% response rate represents the broader population of veterinary students at ULPGC, and what specific characteristics might differentiate responders from non-responders in ways that could systematically bias the findings?
What is the precise total target population from which the 43% response rate was calculated, and what methods were used to verify that the survey invitation reached all eligible students, rather than just those attending specific teaching activities?
In Figure 2(e), why do only 19.9% of students report using AI in permitted online examinations, while 87.1% report using it for individual assignments. Does this reflect institutional policies, assessment format constraints, or students' ethical judgments about appropriate AI use?
How do the authors reconcile Figure 3(c)'s finding that 90%+ of students express concern about misinformation with Figure 5(d)'s finding that 57.9% never or rarely disclose AI use, and what does this discrepancy suggest about the relationship between awareness of AI risks and responsible AI practices?
Figure 4(a) shows 83.6% perceive AI has improved understanding, yet Figure 4(c) shows only 47.4% report acquiring new skills through AI. What explains this discrepancy between perceived content understanding and perceived skill development, and what does it suggest about how students conceptualize learning?
ChatGPT dominated usage (87.1%), while DeepSeek, Claude, and NotebookLM showed minimal adoption (Figure 2c). To what extent does this reflect objective tool superiority versus mere accessibility/visibility, and how might students' preference for ChatGPT over potentially more specialized veterinary AI tools affect their AI literacy development?
The article calls for longitudinal and multi-institutional studies, but what specific hypotheses, outcomes, and comparison groups should such studies investigate to move beyond descriptive findings toward evidence-based curriculum design recommendations?
Can you please explain the relation of your research to the subject of the journal Applied Sciences? Why didn't you send the article to AI in Education (https://www.mdpi.com/journal/aieduc) or Education Sciences (https://www.mdpi.com/journal/education) or Trends in Higher Education (https://www.mdpi.com/journal/higheredu)?

Author Response

We are grateful to the reviewer for the thorough assessment of our manuscript and for the valuable comments and suggestions provided. These remarks have helped us to improve the clarity, quality, and overall scientific strength of the study. We have carefully considered each point raised and have revised the manuscript accordingly. A detailed response to every comment is provided below, together with an explanation of the corresponding modifications introduced in the revised manuscript.

Comment 1: The non-uniform distribution Year 1: 28.6%, Year 2: 12.7%, Year 3: 6.3%, Year 4: 27.5%, Year 5: 24.9% is striking, particularly the underrepresentation of third-year students 6.3%. This raises questions about the representativeness of student perspectives across the curriculum and whether convenience sampling inadvertently excluded certain student populations, e.g., those on clinical rotations. The authors conduct chi-square tests showing significant differences (p<0.001) but do not adequately address how this uneven distribution might affect interpretation of year-specific patterns.

Response: We thank the reviewer for this important observation. We agree that the unequal representation across academic years, particularly the relatively small proportion of third-year students, should be considered when interpreting the results. As the study used convenience sampling during regular teaching activities, differences in student attendance and availability may have contributed to the unequal participation across academic years. However, we did not collect information on non-responders or on the specific reasons for non-participation and therefore cannot determine whether particular groups, such as students undertaking clinical activities, were systematically less likely to participate.

To address the potential impact of this imbalance on the statistical analyses, we did not perform inferential comparisons based on the five individual academic years. Instead, academic year was recategorized into two broader groups, early-stage students, Years 1 and 2, and advanced-stage students, Years 3 to 5. This approach provided more balanced group sizes and reduced sparse cell counts that would have limited the reliability of year-specific comparisons. Associations between academic stage and the questionnaire variables were subsequently reanalysed, including appropriate procedures for low expected cell frequencies and adjustment for multiple comparisons using the Benjamini-Hochberg procedure.

In addition, we have revised the manuscript to explicitly acknowledge that the unequal participation across individual academic years limits the representativeness and interpretation of year-specific patterns. Accordingly, the results are interpreted at the broader academic-stage level rather than as differences between individual years. Moreover, the potential for selection and non-response bias associated with convenience sampling has been expanded in the Study Limitations section.

Comment 2: All data derive from self-report rather than objective assessments of AI literacy, academic performance, clinical reasoning, or critical thinking skills. The authors appropriately acknowledge this limitation, but the discussion frequently treats perceived benefits as if they were demonstrated outcomes. The statement that students reported improvements in conceptual understanding conflates subjective perception with objective learning gain.

Response: We agree that the present study relies on self-reported perceptions and does not include objective assessments of AI literacy, academic performance, clinical reasoning, or critical thinking. We have therefore carefully revised the manuscript to distinguish perceived benefits from objectively demonstrated educational outcomes. In particular, statements referring to improvements in learning or understanding have been rephrased to explicitly indicate that these represent students’ perceptions or self-reported experiences. We have also clarified in the Discussion and Limitations that the cross-sectional design and self-reported nature of the data do not allow us to determine whether perceived improvements translate into objectively measurable educational gains.

Comment 3: The authors reference significant differences in demographic distributions but do not conduct association tests between demographics and AI-related outcomes.

Response: Thank you for this important comment. We agree that exploring associations between demographic and academic characteristics and AI-related outcomes provides additional insight into the survey findings. Accordingly, we expanded the statistical analysis in the revised manuscript. However, due to the uneven distribution across academic years, particularly the small number of third-year students, academic year was grouped into academic stage. First and second-year students were classified as early-stage, while third to fifth-year students were classified as advanced-stage. Associations between academic stage and AI-related variables were assessed using categorical association tests, with Cramér’s V to estimate effect size, exact conditional Monte Carlo procedures when required, and Benjamini-Hochberg correction for multiple comparisons. After FDR (False Discovery Rate) adjustment, academic stage showed significant associations with overall AI use, frequency and timing of first use, patterns and purposes of use, perceived improvement in understanding complex concepts, changes in study methods, and acknowledgement of AI use. No significant associations were found for future expectations regarding AI.

Gender-related analyses revealed some nominally significant associations, but none remained significant after Benjamini-Hochberg correction. No significant gender differences were observed in multiple-response questions after controlling the false discovery rate.

The Materials and Methods, Results and Discussion sections have been revised accordingly. We believe these analyses address the reviewer’s concern and provide a more comprehensive assessment of the relationship between students’ demographic and academic characteristics and their AI use, perceptions, attitudes, experiences, and expectations.

Comment 4: The small sample size for AI non-users n=3 precludes meaningful subgroup analysis, yet the authors present percentages (66.7%, 33.3%) that could be misleading due to the tiny denominator.

Response: We thank the reviewer for this important observation. We agree that the very small number of AI non-users (n = 3) precludes meaningful subgroup analysis and that percentages based on such a small denominator may be misleading. We have therefore replaced the percentages with absolute frequencies in the text and revised Figure 2 accordingly. We have also clarified that these observations should be interpreted descriptively and do not permit meaningful subgroup comparisons

Comment 5: Given the convenience sampling strategy and potential self-selection bias, how confident can we be that the 43% response rate represents the broader population of veterinary students at ULPGC, and what specific characteristics might differentiate responders from non-responders in ways that could systematically bias the findings?

Response: We agree that the convenience sampling strategy and voluntary participation limit the extent to which the 43% response rate can be interpreted as representative of the broader population of veterinary students at ULPGC. Although students from all academic years were represented, the study did not collect information from non-responders; therefore, we cannot directly assess whether responders differed systematically from those who did not participate. It is possible that students with greater interest in AI, greater familiarity with digital technologies, stronger academic engagement, or greater awareness of AI-related issues may have been more likely to participate. Conversely, students with limited interest in or exposure to AI may have been less inclined to complete the survey. Such differences could potentially influence estimates of AI awareness, usage patterns, and attitudes. We have therefore clarified in the revised manuscript that the response rate should not be interpreted as evidence of full representativeness and that selection and non-response bias cannot be excluded. Future studies using probability-based or census sampling and collecting information on non-responders, where feasible, would help assess and reduce this potential bias.

Comment 6: What is the precise total target population from which the 43% response rate was calculated, and what methods were used to verify that the survey invitation reached all eligible students, rather than just those attending specific teaching activities?

Response: We thank the reviewer for highlighting this point. The target population comprised approximately 440 undergraduate students enrolled in the Veterinary Medicine degree at the University of Las Palmas de Gran Canaria during the 2025–2026 academic year, of whom 189 completed the questionnaire, corresponding to a response rate of approximately 43%. The survey was disseminated through regular teaching activities across the Veterinary Medicine programme, where students received a brief explanation of the study and were provided with a QR code linking to the online questionnaire. However, because a convenience sampling strategy was used, we cannot confirm that every eligible student was directly exposed to the survey invitation or that all students had an equal probability of participation. We have clarified this point in the revised Methods section and acknowledge the potential for selection and non-response bias as a limitation of the study.

Comment 7: In Figure 2(e), why do only 19.9% of students report using AI in permitted online examinations, while 87.1% report using it for individual assignments. Does this reflect institutional policies, assessment format constraints, or students' ethical judgments about appropriate AI use?

Response: We agree that the substantially lower use of AI in permitted online examinations (19.9%) compared with individual assignments (87.1%) is noteworthy. Our data do not allow us to determine whether this difference is primarily attributable to institutional policies, assessment format, or students' own ethical judgments. However, the pattern may reflect a combination of these factors. Individual assignments may provide greater flexibility for using AI as a learning or cognitive support tool, whereas examinations, even when AI use is explicitly permitted, may involve stricter assessment conditions and greater concerns about academic integrity or the appropriate role of AI. The lower use in permitted online examinations may therefore indicate that students distinguish between AI use for learning and academic support and its use in contexts intended to evaluate individual performance. This interpretation is consistent with our broader findings regarding the ethical dimension of AI use, although the questionnaire did not directly assess students' reasons for using or avoiding AI in different assessment formats. Future studies should examine how institutional policies, assessment design, and students' ethical perceptions interact to shape AI use across different forms of assessment.

Comment 8: How do the authors reconcile Figure 3(c)'s finding that 90%+ of students express concern about misinformation with Figure 5(d)'s finding that 57.9% never or rarely disclose AI use, and what does this discrepancy suggest about the relationship between awareness of AI risks and responsible AI practices?

Response: We agree that the discrepancy between the high proportion of students expressing concern about misinformation (>90%) and the 57.9% who reported never or rarely disclosing their use of AI is noteworthy. This finding may indicate that awareness of AI-related risks does not necessarily translate into consistent responsible AI practices. Students may recognize the potential for inaccurate or misleading AI-generated information while having less awareness of, or placing less emphasis on, transparency regarding AI use. This distinction is particularly relevant because responsible AI use involves not only recognizing potential risks, but also applying appropriate practices such as verifying AI-generated information, acknowledging AI assistance when appropriate, and critically evaluating AI outputs. The observed discrepancy therefore suggests that AI risk awareness and responsible AI behaviour may represent related but distinct dimensions of AI literacy. However, because our questionnaire did not directly assess the reasons underlying students’ disclosure practices, this interpretation should be considered cautiously. Further research should examine the factors that facilitate the translation of AI risk awareness into consistent and transparent practices among veterinary students.

Comment 9: Figure 4(a) shows 83.6% perceive AI has improved understanding, yet Figure 4(c) shows only 47.4% report acquiring new skills through AI. What explains this discrepancy between perceived content understanding and perceived skill development, and what does it suggest about how students conceptualize learning?

Response: We agree that the difference between perceived improvement in understanding (83.6%) and the acquisition of new skills (47.4%) is noteworthy. This discrepancy may reflect a distinction between using AI as a tool for supporting content comprehension and using it as a means of developing broader skills or competencies. In our study, students primarily reported using AI for answering questions, searching for information, exam preparation, and brainstorming, suggesting that AI was mainly perceived as a cognitive support tool rather than as a substitute for active skill development. Thus, students may perceive that AI helps them understand or clarify existing knowledge without necessarily contributing to the acquisition of new practical, analytical, or transferable skills. This finding may also suggest that students conceptualize learning primarily in terms of knowledge acquisition and understanding, rather than competency development. However, because our cross-sectional questionnaire did not directly assess how students define or conceptualize learning, this interpretation should be considered cautiously. Future studies should investigate how students distinguish between knowledge acquisition, conceptual understanding, and skill development when using AI in educational settings.

Comment 10: ChatGPT dominated usage (87.1%), while DeepSeek, Claude, and NotebookLM showed minimal adoption (Figure 2c). To what extent does this reflect objective tool superiority versus mere accessibility/visibility, and how might students' preference for ChatGPT over potentially more specialized veterinary AI tools affect their AI literacy development?

Response: We agree that our data do not allow us to determine whether ChatGPT's predominance reflects superior performance or simply greater accessibility, visibility, and familiarity among students. The latter factors may have played an important role, given ChatGPT's widespread public exposure and ease of access. The preference for a general-purpose AI tool may also have implications for AI literacy development, as students may become proficient users of a single platform without necessarily developing the skills needed to critically compare different AI systems, assess their limitations, or select the most appropriate tool for specific professional tasks. Future studies should therefore explore how the use of different AI platforms influences AI literacy and critical evaluation skills.

Comment 11: The article calls for longitudinal and multi-institutional studies, but what specific hypotheses, outcomes, and comparison groups should such studies investigate to move beyond descriptive findings toward evidence-based curriculum design recommendations?

Response: We thank the reviewer for this suggestion. We agree that future studies should move beyond descriptive findings and evaluate objective outcomes. To clarify this point, we have expanded the Conclusions section by specifying potential comparison groups, including students receiving structured AI literacy training versus those following standard curricula across different institutions and academic years.

Comment 12: Can you please explain the relation of your research to the subject of the journal Applied Sciences? Why didn't you send the article to AI in Education (https://www.mdpi.com/journal/aieduc) or Education Sciences (https://www.mdpi.com/journal/education) or Trends in Higher Education (https://www.mdpi.com/journal/higheredu)?

Response: Thank you for pointing this out. Although the study was conducted in a higher education setting, its primary focus is AI adoption, AI literacy, and the responsible use of AI technologies by future veterinary professionals rather than educational theory or pedagogical interventions. The manuscript addresses the practical implications of AI implementation in a professional field where AI is increasingly applied in diagnostics, clinical decision-making, and animal health. For this reason, we believe the article fits well within the scope of Applied Sciences. We also note that the submission followed an invitation from the journal's Editorial Office.

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript focuses on the gap between AI adoption and AI literacy in veterinary education, based on a cross-sectional survey of 189 veterinary students from ULPGC. This research has practical significance, but there are also areas that need further explanation and clarification as follows.
1.Both the title and the core conclusions of this paper use the term "AI literacy," but the questionnaire mainly measures dimensions including students' knowledge, awareness, usage patterns, perceptions, and attitudes. The results merely reflect students' subjective judgments and do not prove their actual AI literacy levels. Although the authors acknowledge the lack of objective AI literacy measurements in the Limitations section, if supplementary measurements cannot be added, it is recommended that the Methods section clarify that this research only evaluates perceived literacy rather than actual literacy.
2.The questionnaire recruited a sample of 189 ULPGC veterinary students. The overall sample size is acceptable, but the distribution across academic years is notably uneven. First‑year (28.6%) and fourth‑year (27.5%) students constitute the majority, whereas third‑year students account for only 6.3%. To some extent, the third year is typically the most intensive stage in most veterinary curricula, and the under‑representation of this group may introduce bias. Therefore, I would recommend that the authors conduct a targeted supplementary survey to collect additional data from third‑year students, thereby achieving a more balanced distribution across year groups.
3.Since the manuscript collected data ranging from the first to the fifth academic year, covering different educational stages, I suggest the authors conduct Chi-square tests to analyze the differences in AI attitudes and usage among students of different academic years in depth. This would significantly enhance the research depth of the paper.
4. The sample is drawn from a single source—only one university in Spain. This is a significant weakness for a questionnaire survey, as it limits the generalizability of the findings to other regions or countries. Although the authors have mentioned this shortcoming, I recommend they elaborate more fully on the limitations caused by this sampling bias.
5.Given the significant shortcomings in the survey sample mentioned above, how can the paper justify its claim of providing a "robust institutional overview"? Please clarify and support this statement.
6.In the "Future expectations" section, I suggest the authors closely tie their proposed future directions to the specific limitations identified in this study, particularly those related to the sample and methodology.

Comments on the Quality of English Language

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

Author Response

We would like to express our sincere appreciation to the reviewer for the time and effort devoted to evaluating our manuscript. The insightful comments and recommendations have been highly valuable in strengthening the manuscript and enhancing its scientific quality. We have addressed all concerns raised and implemented the necessary revisions. Detailed responses to each comment, along with the relevant changes made in the manuscript, are presented below.

Comment 1: Both the title and the core conclusions of this paper use the term "AI literacy," but the questionnaire mainly measures dimensions including students' knowledge, awareness, usage patterns, perceptions, and attitudes. The results merely reflect students' subjective judgments and do not prove their actual AI literacy levels. Although the authors acknowledge the lack of objective AI literacy measurements in the Limitations section, if supplementary measurements cannot be added, it is recommended that the Methods section clarify that this research only evaluates perceived literacy rather than actual literacy.

Response: We thank the reviewer for this important clarification. We agree that our questionnaire assessed students’ self-reported knowledge, awareness, use, perceptions, and attitudes toward AI rather than providing an objective or validated measure of AI literacy. We have therefore revised the Methods section to explicitly state that the study evaluates perceived knowledge and awareness of AI and self-reported engagement with AI tools, rather than objectively measured AI literacy competencies. We have also reviewed the manuscript to avoid interpreting these self-reported measures as evidence of objectively established AI literacy. Where appropriate, statements referring to AI literacy have been rephrased to emphasize perceived AI literacy, preparedness, or the need for AI literacy development. We have additionally revised the title and relevant conclusions to ensure that they accurately reflect the scope of the questionnaire and do not imply that AI literacy was objectively measured.

Comment 2: The questionnaire recruited a sample of 189 ULPGC veterinary students. The overall sample size is acceptable, but the distribution across academic years is notably uneven. First‑year (28.6%) and fourth‑year (27.5%) students constitute the majority, whereas third‑year students account for only 6.3%. To some extent, the third year is typically the most intensive stage in most veterinary curricula, and the under‑representation of this group may introduce bias. Therefore, I would recommend that the authors conduct a targeted supplementary survey to collect additional data from third‑year students, thereby achieving a more balanced distribution across year groups.

Response: We thank the reviewer for highlighting the uneven distribution of participants across academic years and, in particular, the relatively low representation of third-year students. We agree that this imbalance should be considered when interpreting comparisons according to academic year. However, we respectfully consider that conducting a targeted supplementary survey of third-year students would not be methodologically appropriate within the design of the present study. This investigation was designed as a cross-sectional survey, with data collected during a predefined period between February and March 2026. Recruitment was conducted during regular teaching activities using the same procedures for students across the Veterinary Medicine programme. The original data collection period has now concluded, and students are currently outside the regular teaching period. Reopening recruitment specifically for third-year students at a different time and under different recruitment conditions would therefore introduce a second data collection period and a targeted sampling procedure that would differ from the original study protocol. Given the rapid evolution of generative AI technologies and their use in higher education, responses obtained at a later time could also reflect a different exposure context.

Indeed, the uneven distribution across academic years was one of the reasons why, in the original analysis, we deliberately did not perform comparisons between responses to the questionnaire and individual academic years. In particular, the small number of third-year respondents (n = 12) would have resulted in comparisons based on markedly unequal subgroup sizes. We therefore considered that presenting year-specific inferential analyses could lead to unstable estimates and potentially biased interpretation of differences between academic years. The original results were consequently presented primarily for the overall sample rather than emphasizing comparisons among the five individual academic years.

To address the reviewer's concern while preserving the original cross-sectional design, we have taken two additional steps. First, we have explicitly acknowledged the uneven distribution across academic years, and particularly the under-representation of third-year students, as a limitation of the study. The revised manuscript now emphasizes that this imbalance may limit the interpretation of year-specific patterns.

Second, following the reviewer's comment and other reviewers' suggestions concerning subgroup analyses, we performed additional analyses according to stage of academic training. Rather than comparing all five academic years separately, students were grouped into early-stage students (first and second years) and later-stage students (third, fourth, and fifth years). This approach allowed us to examine potential differences according to a broader stage of veterinary training while avoiding interpretation based on the particularly small third-year subgroup. Associations were assessed using Pearson's chi-square test, with Cramér's V used to estimate effect size. The Benjamini-Hochberg procedure was applied to control the false discovery rate arising from multiple comparisons.

The corresponding results have now been incorporated into the revised manuscript. Importantly, this additional analysis should be considered exploratory and does not overcome the limitations associated with convenience sampling or the uneven representation of the individual academic years. We have therefore retained this issue as an explicit limitation and have avoided interpreting the findings as representative of each academic year independently.

Comment 3: Since the manuscript collected data ranging from the first to the fifth academic year, covering different educational stages, I suggest the authors conduct Chi-square tests to analyze the differences in AI attitudes and usage among students of different academic years in depth. This would significantly enhance the research depth of the paper.

Response: Thank you for this comment. As this concern largely overlaps with the previous reviewer’s comment regarding associations between demographic and academic characteristics and AI-related outcomes, we provide the same response below.

We agree that exploring associations between demographic and academic characteristics and AI-related outcomes provides additional insight into the survey findings. Accordingly, we expanded the statistical analysis in the revised manuscript. However, due to the uneven distribution across academic years, particularly the small number of third-year students, academic year was grouped into academic stage. First and second-year students were classified as early-stage, while third to fifth-year students were classified as advanced-stage. Associations between academic stage and AI-related variables were assessed using categorical association tests, with Cramér’s V to estimate effect size, exact conditional Monte Carlo procedures when required, and Benjamini-Hochberg correction for multiple comparisons. After FDR (False Discovery Rate) adjustment, academic stage showed significant associations with overall AI use, frequency and timing of first use, patterns and purposes of use, perceived improvement in understanding complex concepts, changes in study methods, and acknowledgement of AI use. No significant associations were found for future expectations regarding AI.

Comment 4: The sample is drawn from a single source—only one university in Spain. This is a significant weakness for a questionnaire survey, as it limits the generalizability of the findings to other regions or countries. Although the authors have mentioned this shortcoming, I recommend they elaborate more fully on the limitations caused by this sampling bias.

Response: We thank the reviewer for highlighting this important limitation. We agree that the single-institution setting represents an important constraint on the generalizability of our findings. We have expanded the Limitations section to clarify that differences in curriculum structure, institutional AI policies, availability of AI training and digital resources, student characteristics, and cultural and regulatory contexts may influence students’ knowledge, use, and perceptions of AI. We now explicitly describe the findings as an institution-specific characterization rather than as representative of veterinary students more broadly. We have also emphasized the need for multi-institutional studies involving veterinary schools from different regions and educational contexts to assess the consistency and generalizability of the observed patterns.

Comment 5: Given the significant shortcomings in the survey sample mentioned above, how can the paper justify its claim of providing a "robust institutional overview"? Please clarify and support this statement.

Response: We thank the reviewer for pointing out this issue. We agree that the term “robust institutional overview” was too strong given the voluntary convenience sampling, the uneven representation across academic years, the potential for non-response bias, and the lack of information about non-participants. Although the survey included students from all academic years and achieved a 43% response rate among the target population, these characteristics do not allow us to claim that the sample provides a fully representative institutional overview. We have therefore removed this wording and revised the corresponding statements to describe the study more cautiously as providing an institution-specific overview of veterinary students’ reported knowledge, use, perceptions, and attitudes toward AI. We have also clarified the relevant sampling limitations in the Limitations section.

Comment 6: In the "Future expectations" section, I suggest the authors closely tie their proposed future directions to the specific limitations identified in this study, particularly those related to the sample and methodology.

Response: Thank you for pointing this out. We agree that the future research directions should be more explicitly connected to the limitations identified in the present study. We have therefore revised the “Future expectations” section to link specific methodological and sampling limitations with corresponding research priorities. In particular, the revised text now highlights the need for multi-institutional studies using more systematic or stratified sampling to address the single-institution setting, uneven representation across academic years, and potential non-response bias; longitudinal studies and objective measures of AI literacy and educational outcomes to address the cross-sectional and self-reported nature of the present study; and the development and validation of standardized instruments to facilitate more robust assessment and comparison of AI literacy among veterinary students.

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I would like to thank the authors for revising the manuscript and for taking the previous comments into account. The manuscript has improved considerably. However, several issues still need to be addressed before it can be considered for publication.

I still strongly recommend including a separate Related Work section and conducting a more comprehensive search and review of the relevant literature. From my point of view, the literature on AI adaptation in veterinary education is not as limited as the authors indicate. Few examples of the recent publications that are not included in the review given in the paper are the following:

Oropesa A. L., et al. (2026). Artificial intelligence in veterinary education: self-perceived knowledge, use, and attitudes among veterinary students in Spain and Portugal...
Selcuk A., et al. (2026). Artificial intelligence in veterinary education: self-perceived knowledge, use, and attitudes among veterinary students in Spain and Portugal...
Li S., et al. (2026). The adoption paradox for veterinary professionals in China: high use of artificial intelligence despite low familiarity...

Unfortunately, the authors have still not clearly formulated the research questions. Explicit research questions are an essential component of the study because they provide the foundation for the development of the research instruments, the study design, the analytical approach, and the interpretation of the results. I strongly recommend stating the research questions explicitly and ensuring that the methodology and results are clearly aligned with them.

In addition, the following editorial and formatting issues should be addressed:

The first-person possessive pronoun “our” is still used in the manuscript.
I recommend using the full terms rather than abbreviations in the list of keywords.
The authors introduced the abbreviation “FDR” in the revised text; however, on p. 12, the full term is used again. Please ensure that abbreviations are introduced and used consistently throughout the manuscript.
The text in several figures is difficult to read, as many letters appear to overlap or be positioned too closely together. Please improve the resolution, formatting, and legibility of all figures.
Please use consistent spacing between citation numbers when multiple references are cited together. For example, write [2, 3, 4] rather than [2,3,4].

Author Response

We sincerely thank the reviewer for their thorough evaluation of the revised manuscript and for recognizing the substantial improvements made in response to the previous round of comments. We greatly appreciate the reviewer’s constructive feedback, careful reassessment of the work, and continued engagement with the manuscript. We also acknowledge that several concerns remain unresolved and agree that these issues should be adequately addressed before the manuscript can be considered for publication. Below, we provide a detailed point-by-point response to each comment.

Comment 1: I still strongly recommend including a separate Related Work section and conducting a more comprehensive search and review of the relevant literature. From my point of view, the literature on AI adaptation in veterinary education is not as limited as the authors indicate. Few examples of the recent publications that are not included in the review given in the paper are the following:

Oropesa A. L., et al. (2026). Artificial intelligence in veterinary education: self-perceived knowledge, use, and attitudes among veterinary students in Spain and Portugal...
Selcuk A., et al. (2026). Artificial intelligence in veterinary education: self-perceived knowledge, use, and attitudes among veterinary students in Spain and Portugal...
Li S., et al. (2026). The adoption paradox for veterinary professionals in China: high use of artificial intelligence despite low familiarity...

Response: We sincerely thank the reviewer for this valuable suggestion and for highlighting additional relevant literature in the field. We agree that the growing body of research on artificial intelligence in veterinary education and professional practice deserves a more comprehensive discussion than was provided in the previous version of the manuscript.

In response to this comment, we have now incorporated a dedicated Related Work section into the manuscript, where we contextualize our study within the broader literature on AI adoption, perceptions, and educational applications in veterinary medicine. We have also expanded the literature review through an updated search and have included the references recommended by the reviewer, including the studies by Oropesa et al. (2026), and Li et al. (2026). These publications are now discussed in relation to our findings and are cited appropriately in the revised manuscript.

We believe that these additions have strengthened the theoretical background of the study, provided a more comprehensive overview of current research in the field, and clarified the contribution of our work within the existing literature. We are grateful to the reviewer for this constructive recommendation, which has significantly improved the manuscript.

Comment 2: Unfortunately, the authors have still not clearly formulated the research questions. Explicit research questions are an essential component of the study because they provide the foundation for the development of the research instruments, the study design, the analytical approach, and the interpretation of the results. I strongly recommend stating the research questions explicitly and ensuring that the methodology and results are clearly aligned with them.

Response: Thank you for this valuable comment. We agree that clearly stated research questions are essential for guiding the study design, methodology, analysis, and interpretation of results. In response to your suggestion, we have revised the Introduction to explicitly formulate the research questions and objectives of the study. We have also reviewed the manuscript to ensure that the methodology, results, and discussion are clearly aligned with these research questions, thereby improving the overall coherence and transparency of the research framework.

Comment 3: The first-person possessive pronoun “our” is still used in the manuscript.

Response: We thank the reviewer for this careful observation. In response to this comment, we have thoroughly reviewed the manuscript and removed the remaining instances of the first-person possessive pronoun “our” wherever appropriate. The corresponding sentences have been revised to maintain a more objective and impersonal academic style, consistent with the journal’s writing conventions.

Comment 4: I recommend using the full terms rather than abbreviations in the list of keywords.

Response: We thank the reviewer for this helpful suggestion. We agree that using full terms in the keyword list improves clarity, accessibility, and discoverability of the manuscript. In response to this comment, we have revised the keywords section and replaced all abbreviations with their corresponding full terms throughout the list.

Comment 5: The authors introduced the abbreviation “FDR” in the revised text; however, on p. 12, the full term is used again. Please ensure that abbreviations are introduced and used consistently throughout the manuscript.

Response: We thank the reviewer for pointing out this inconsistency. In response to this comment, we have carefully reviewed the manuscript and ensured that the abbreviation FDR (False Discovery Rate) is introduced once at its first occurrence and then used consistently throughout the text. The instance on page 12 where the full term was repeated has been revised accordingly. We appreciate the reviewer’s attention to detail, which has helped improve the consistency and readability of the manuscript.

Comment 6: The text in several figures is difficult to read, as many letters appear to overlap or be positioned too closely together. Please improve the resolution, formatting, and legibility of all figures.

Response: Thank you for this observation. We carefully revised all figures to improve their readability and overall presentation. In the figures where text overlap or insufficient spacing occurred (Figures 2 and 3), the affected elements were repositioned to eliminate overlaps and improve legibility. In addition, Figure 1c was reformatted to ensure a more uniform distribution of elements. The figures were also exported at high resolution to enhance clarity and readability throughout the manuscript.

Comment 7: Please use consistent spacing between citation numbers when multiple references are cited together. For example, write [2, 3, 4] rather than [2,3,4].

Response: Thank you for this observation. We have revised the manuscript to ensure consistent spacing between citation numbers throughout the text. All multiple citations are now formatted according to the journal style, e.g., [2, 3, 4] instead of [2,3,4].

Reviewer 2 Report

Comments and Suggestions for Authors

The reviewer thanks the authors for the good answers to all the comments. The manuscript has been thoroughly revised. I propose to accept the article in its present form.

Author Response

We sincerely thank the reviewer for the positive assessment of our manuscript and for the kind acknowledgment of our responses and revisions. We greatly appreciate the reviewer’s careful evaluation and valuable comments throughout the review process. We are particularly grateful for the recommendation to accept the manuscript in its present form.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have addressed all of my comments, and I have no further concerns. The manuscript is now suitable for publication.

Comments on the Quality of English Language The English can be improved to express the research more clearly.

Author Response

Thank you again for your valuable comments and suggestions. We appreciate the reviewer’s observation regarding the clarity and quality of the English language. In response to this suggestion, we have carefully revised the manuscript to improve the English language and enhance the clarity, precision, and readability of the presentation throughout the text.