Enabling and Inhibitory Pathways of University Students' Willingness to Disclose AI Use: A Cognition-Affect-Conation Perspective

Huimin He; Yiran Du

arxiv: 2604.21733 · v1 · submitted 2026-04-23 · 💻 cs.AI

Enabling and Inhibitory Pathways of University Students' Willingness to Disclose AI Use: A Cognition-Affect-Conation Perspective

Yiran Du , Huimin He This is my paper

Pith reviewed 2026-05-09 22:06 UTC · model grok-4.3

classification 💻 cs.AI

keywords AI use disclosurepsychological safetyevaluation apprehensionhigher educationCognition-Affect-Conation frameworkmixed methodsstudent transparency

0 comments

The pith

Psychological safety encourages university students to disclose AI use in their work, while evaluation apprehension discourages it.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The study applies the Cognition-Affect-Conation framework to map psychological routes to students' decisions about revealing AI assistance in academic tasks. Quantitative modeling of survey responses from 546 students shows that perceptions of fairness, teacher support, and organizational backing build psychological safety, which raises disclosure willingness. At the same time, feelings of stigma, uncertainty, and privacy worry heighten evaluation apprehension, which lowers both safety and disclosure intention. Follow-up interviews with 22 students add that institutional clarity fosters openness while policy vagueness prompts cautious or selective reporting strategies.

Core claim

Applying the CAC framework, the research establishes that psychological safety acts as an enabling pathway shaped by perceived fairness, teacher support, and organizational support to increase willingness to disclose AI use, whereas evaluation apprehension serves as an inhibitory pathway strengthened by perceived stigma, uncertainty, and privacy concern to reduce disclosure intention.

What carries the argument

Cognition-Affect-Conation (CAC) framework, where cognitive perceptions shape affective states of psychological safety and evaluation apprehension that then drive conative disclosure intention.

If this is right

Clear institutional policies and supportive teaching practices increase students' openness about AI use.
Perceived fairness and organizational backing strengthen the psychological safety that promotes disclosure.
Reducing perceived stigma and uncertainty through guidance lowers evaluation apprehension and its negative effects.
Ambiguous policies lead students to adopt strategic or partial disclosure approaches instead of full transparency.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Universities could revise AI guidelines to emphasize support and clarity over potential penalties to raise overall transparency rates.
The enabling and inhibitory pathways identified here may operate similarly when students consider disclosing other emerging tools or in workplace contexts.
Longitudinal tracking of actual submission behaviors could test whether the modeled intentions translate into sustained changes in reporting habits.

Load-bearing premise

Self-reported survey responses and interview statements accurately reflect students' actual disclosure behaviors and underlying psychological states without significant social desirability bias or measurement error in the CAC constructs.

What would settle it

A field study that directly measures whether students actually disclose AI use in submitted assignments under manipulated conditions of institutional support versus stigma, then compares those rates to their self-reported intentions.

Figures

Figures reproduced from arXiv: 2604.21733 by Huimin He, Yiran Du.

read the original abstract

The increasing integration of artificial intelligence (AI) in higher education has raised important questions regarding students' transparency in reporting AI-assisted work. This study investigates the psychological mechanisms underlying university students' willingness to disclose AI use by applying the Cognition--Affect--Conation (CAC) framework. A sequential explanatory mixed-methods design was employed. In the quantitative phase, survey data were collected from 546 university students and analysed using structural equation modelling to examine the relationships among cognitive perceptions, affective responses, and disclosure intention. In the qualitative phase, semi-structured interviews with 22 students were conducted to further interpret the quantitative findings. The results indicate that psychological safety significantly increases students' willingness to disclose AI use and is positively shaped by perceived fairness, perceived teacher support, and perceived organisational support. Conversely, evaluation apprehension reduces disclosure intention and psychological safety, and is strengthened by perceived stigma, perceived uncertainty, and privacy concern. Qualitative findings further reveal that institutional clarity and supportive instructional practices encourage openness, whereas policy ambiguity and fear of negative evaluation often lead students to adopt cautious or strategic disclosure practices. Overall, the study highlights the dual role of enabling and inhibitory psychological mechanisms in shaping AI-use disclosure and underscores the importance of supportive institutional environments and clear guidance for promoting responsible AI transparency in higher education.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper maps CAC pathways to AI disclosure via self-reports but lacks behavioral checks or SEM diagnostics, leaving the claims tentative.

read the letter

The main point is that this study uses the cognition-affect-conation framework to trace how perceived fairness, teacher support, and organisational support build psychological safety that raises students' willingness to disclose AI use, while stigma, uncertainty, and privacy concerns heighten evaluation apprehension that lowers it. The 546-student survey plus 22 interviews add some interpretive depth to those paths. What is new is the concrete identification of these enabling and inhibitory routes in the specific setting of university AI transparency, which prior work cited in the abstract had not covered. The paper handles the mixed-methods sequence cleanly enough to link the quantitative model to student comments on institutional clarity versus policy ambiguity. The soft spots sit mainly in the evidence base. The abstract gives no model fit numbers, convergent validity checks, or common-method bias tests for the SEM, and the full reliance on self-reported intentions in a high-stakes academic context invites social-desirability bias. Without any cross-check against actual submission logs or experimental tasks, it is hard to know how much the reported mechanisms reflect real behavior rather than stated norms. This work is aimed at higher-education researchers and policy staff who deal with academic integrity and AI adoption. Readers looking for practical levers around support and clarity could extract usable ideas from the qualitative section. It deserves a serious referee because the topic is current and the theoretical framing is straightforward, even though the measurement limitations would require attention in revision.

Referee Report

3 major / 2 minor

Summary. The manuscript applies the Cognition-Affect-Conation (CAC) framework in a sequential explanatory mixed-methods study to investigate university students' willingness to disclose AI use in academic work. Quantitative data from 546 survey responses are analyzed via structural equation modeling to test paths from cognitive perceptions (perceived fairness, teacher support, organizational support, stigma, uncertainty, privacy concern) through affective responses (psychological safety, evaluation apprehension) to disclosure intention. This is followed by thematic analysis of semi-structured interviews with 22 students to interpret the quantitative results. The central findings are that psychological safety positively predicts disclosure willingness and is enhanced by fairness, teacher support, and organizational support, while evaluation apprehension negatively predicts disclosure willingness and is strengthened by stigma, uncertainty, and privacy concern; qualitative data highlight the roles of institutional clarity versus policy ambiguity.

Significance. If the SEM results prove robust after standard validation, the study would offer a theoretically grounded account of enabling and inhibitory pathways for AI transparency in higher education, with practical implications for institutional policies and teaching practices. The sequential mixed-methods design is a methodological strength that allows quantitative patterns to be contextualized qualitatively. However, the current lack of reported model diagnostics and behavioral validation substantially reduces the immediate contribution and generalizability of the claims.

major comments (3)

[Quantitative phase / SEM analysis] The abstract and quantitative phase description report SEM results on the 546 responses but provide no goodness-of-fit statistics (CFI, RMSEA, SRMR, χ²/df), reliability coefficients, convergent validity (AVE), or discriminant validity metrics. These diagnostics are required to establish that the measurement model supports the structural paths (e.g., psychological safety → disclosure intention) that constitute the paper's central claim.
[Methods / Data collection] All constructs are measured via self-report Likert scales collected from the same participants. No test for common method bias (Harman's single-factor test, marker variable, or latent method factor) is described. In a sensitive topic where non-disclosure can carry academic sanctions, this omission directly threatens the validity of the reported CAC relationships.
[Discussion / Limitations] The study measures and interprets self-reported disclosure willingness and psychological states but reports no objective behavioral indicators (submission logs, AI-detection flags, or experimental tasks). The weakest assumption—that stated intentions accurately reflect underlying CAC mechanisms without substantial social desirability bias—remains unaddressed and is load-bearing for the applied conclusions.

minor comments (2)

[Abstract] The abstract states a 'sequential explanatory' design but does not indicate which specific quantitative findings were selected for probing in the 22 interviews or how integration occurred.
[Quantitative phase] No sample demographics (gender, academic year, discipline, institution type) are summarized, limiting assessment of generalizability even though the quantitative sample size is reported.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their detailed and constructive feedback on our manuscript. The comments highlight important areas for strengthening the methodological transparency and acknowledging limitations. We address each major comment point by point below, outlining specific revisions we will implement.

read point-by-point responses

Referee: [Quantitative phase / SEM analysis] The abstract and quantitative phase description report SEM results on the 546 responses but provide no goodness-of-fit statistics (CFI, RMSEA, SRMR, χ²/df), reliability coefficients, convergent validity (AVE), or discriminant validity metrics. These diagnostics are required to establish that the measurement model supports the structural paths (e.g., psychological safety → disclosure intention) that constitute the paper's central claim.

Authors: We appreciate the referee's emphasis on rigorous reporting of SEM diagnostics. While the manuscript reports basic reliability coefficients (Cronbach's alpha) for the constructs, we acknowledge that a complete set of model fit indices, convergent validity (AVE), and discriminant validity metrics was not presented in the results section. In the revised manuscript, we will add a dedicated measurement model assessment subsection, including goodness-of-fit statistics (CFI, RMSEA, SRMR, χ²/df), factor loadings, composite reliability, AVE values, and discriminant validity via the Fornell-Larcker criterion. These additions will directly support the validity of the reported structural paths. revision: yes
Referee: [Methods / Data collection] All constructs are measured via self-report Likert scales collected from the same participants. No test for common method bias (Harman's single-factor test, marker variable, or latent method factor) is described. In a sensitive topic where non-disclosure can carry academic sanctions, this omission directly threatens the validity of the reported CAC relationships.

Authors: We agree that common method bias poses a significant threat in single-source self-report designs, particularly for sensitive academic behaviors. Although not included in the original submission, we will conduct Harman's single-factor test on the dataset in the revision and report the results (e.g., variance explained by the first unrotated factor). If the test suggests minimal bias, we will include this as evidence of robustness; otherwise, we will discuss it explicitly as a limitation and note potential implications for the CAC pathways. revision: yes
Referee: [Discussion / Limitations] The study measures and interprets self-reported disclosure willingness and psychological states but reports no objective behavioral indicators (submission logs, AI-detection flags, or experimental tasks). The weakest assumption—that stated intentions accurately reflect underlying CAC mechanisms without substantial social desirability bias—remains unaddressed and is load-bearing for the applied conclusions.

Authors: The referee correctly identifies a key limitation of our survey-based design: the absence of objective behavioral measures to validate self-reported intentions. Due to ethical constraints around academic sanctions and the practical challenges of accessing submission logs or detection data, such indicators were not collected. In the revised manuscript, we will expand the limitations section to explicitly address the potential gap between stated willingness and actual behavior, including risks of social desirability bias. We will also outline future research avenues, such as experimental tasks or institutional data analysis, to test these mechanisms behaviorally. The qualitative interviews offer some contextual triangulation, but we recognize this does not fully resolve the issue. revision: partial

Circularity Check

0 steps flagged

No circularity: empirical SEM and interview analysis on new survey data

full rationale

The paper reports a sequential explanatory mixed-methods study that collects fresh survey responses from 546 students and conducts 22 interviews, then applies standard structural equation modelling and thematic coding to test CAC-framework relationships. No equations, fitted parameters, or first-principles derivations are presented that could reduce to their own inputs by construction. All load-bearing claims rest on the external data rather than self-definition, self-citation chains, or renaming of prior results. The analysis is therefore self-contained against its own collected observations.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The study rests on standard assumptions of psychological measurement and statistical modeling rather than introducing new free parameters, axioms, or entities.

axioms (1)

domain assumption Standard assumptions of structural equation modeling hold, including multivariate normality, correct model specification, and absence of omitted variable bias.
Implicit in the use of SEM to test the CAC pathways.

pith-pipeline@v0.9.0 · 5529 in / 1212 out tokens · 28480 ms · 2026-05-09T22:06:04.703686+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

14 extracted references · 14 canonical work pages

[1]

Introduction Artificial intelligence (AI) technologies are increasingly integrated into higher education, transforming how students access information, generate content, and complete academic tasks. Tools based on generative AI and large language models enable students to obtain explanations, draft written material, and support problem-solving activities,...

work page 2023
[2]

Literature Review 2.1 AI in Higher Education Artificial intelligence (AI) technologies have rapidly expanded within higher education, reshaping teaching, learning, and assessment practices (Qian, 2025). Tools based on natural language processing, machine learning, and generative models increasingly support activities such as automated feedback, personalis...

work page 2025
[3]

Theoretical Framework This study draws on the Cognition–Affect–Conation (CAC) framework to explain university students’ willingness to disclose their use of AI in academic contexts. The CAC framework conceptualises behaviour as a sequential psychological process in which individuals’ cognitive evaluations influence their affective responses, which subsequ...

work page 2025
[4]

Research Questions and Hypothesis Development 4.1 Enabling Pathway of Willingness to Disclose AI Use Based on the conceptual model, the first research question (RQ1) asks how psychological safety, perceived fairness, perceived teacher support, and perceived organisational support influence university students’ willingness to disclose AI use. Psychological...

work page 2023
[5]

Methods 5.1 Research Design This study adopted a sequential explanatory mixed-methods design, in which quantitative data collection and analysis were followed by qualitative inquiry to further interpret the quantitative findings (Cohen et al., 2018). In the first phase, a questionnaire survey was administered to examine the relationships among the constru...

work page 2018
[6]

Discussion 7.1 Enabling Pathway of Willingness to Disclose AI Use The quantitative findings provide clear support for the enabling pathway proposed in the Cognition–Affect–Conation framework, demonstrating that psychological safety is a central mechanism shaping students’ willingness to disclose AI use. Structural modelling showed that psychological safet...

work page 2024
[7]

The findings show that psychological safety plays a central role in encouraging disclosure, shaped by students’ perceptions of fairness, teacher support, and organisational support

Conclusion This study examined university students’ willingness to disclose AI use through the Cognition–Affect–Conation framework by identifying both enabling and inhibitory psychological pathways. The findings show that psychological safety plays a central role in encouraging disclosure, shaped by students’ perceptions of fairness, teacher support, and ...

work page doi:10.1108/ijis-08-2022-0165 2025
[8]

watching-eye

https://doi.org/10.3390/informatics12020036 Doğan, M., Celik, A., & Arslan, H. (2025). AI in higher education: Risks and opportunities from the academician perspective. European Journal of Education, 60(1), e12863. https://doi.org/10.1111/ejed.12863 Du, C., Tang, M., Wang, C., Zou, B., Xia, Y ., & Du, Y . (2025). Who is most likely to accept AI chatbots? ...

work page doi:10.3390/informatics12020036 2025
[9]

https://doi.org/10.1186/s40468-023-00219-3 Jeilani, A., & Abubakar, S. (2025). Perceived institutional support and its effects on student perceptions of AI learning in higher education: The role of mediating perceived learning outcomes and moderating technology self-efficacy. Frontiers in Education, 10, 1548900. https://doi.org/10.3389/feduc.2025.1548900 ...

work page doi:10.1186/s40468-023-00219-3 2025
[10]

Z., & Belton, B

https://doi.org/10.1186/s40359-025-02412-w Prasad, K., Aladini, A., Normurodova, N. Z., & Belton, B. (2023). Take language assessment easy: The mediator impacts of self-assessment, test-taking skills in predicting student evaluation apprehension, foreign language learning self-esteem, and language achievement in online classes. Language Testing in Asia, 13(1),

work page doi:10.1186/s40359-025-02412-w 2023
[11]

https://doi.org/10.1186/s40468-023-00246-0 Qian, Y . (2025). Pedagogical applications of generative AI in higher education: A systematic review of the field. TechTrends, 69(5), 1105–1120. https://doi.org/10.1007/s11528-025-01100-1 Somerville, S. G., Harrison, N. M., & Lewis, S. A. (2023). Twelve tips for the pre-brief to promote psychological safety in si...

work page doi:10.1186/s40468-023-00246-0 2025
[12]

https://doi.org/10.3390/ijerph19127499 Usher, M., & Barak, M. (2024). Unpacking the role of AI ethics online education for science and engineering students. International Journal of STEM Education, 11(1),

work page doi:10.3390/ijerph19127499 2024
[13]

https://doi.org/10.1186/s40594-024-00493-4 Usman, M., Cheng, J., Ghani, U., Gul, H., & Shah, W. U. (2021). Social support and perceived uncertainties during COVID-19: Consequences for employees’ wellbeing. Current Psychology. https://doi.org/10.1007/s12144-021-02293-3 Vella, S. A., Mayland, E., Schweickle, M. J., Sutcliffe, J. T., McEwan, D., & Swann, C. ...

work page doi:10.1186/s40594-024-00493-4 2021
[14]

https://doi.org/10.3390/jtaer20020099 Zhou, T., & Wu, X. (2025b). Examining generative AI user disclosure intention: An ELM perspective. Universal Access in the Information Society, 24(2), 1209–1220. https://doi.org/10.1007/s10209-024-01130-1 Zhou, T., & Zhang, C. (2024). Examining generative AI user addiction from a C-A-C perspective. Technology in Socie...

work page doi:10.3390/jtaer20020099 2024

[1] [1]

Introduction Artificial intelligence (AI) technologies are increasingly integrated into higher education, transforming how students access information, generate content, and complete academic tasks. Tools based on generative AI and large language models enable students to obtain explanations, draft written material, and support problem-solving activities,...

work page 2023

[2] [2]

Literature Review 2.1 AI in Higher Education Artificial intelligence (AI) technologies have rapidly expanded within higher education, reshaping teaching, learning, and assessment practices (Qian, 2025). Tools based on natural language processing, machine learning, and generative models increasingly support activities such as automated feedback, personalis...

work page 2025

[3] [3]

Theoretical Framework This study draws on the Cognition–Affect–Conation (CAC) framework to explain university students’ willingness to disclose their use of AI in academic contexts. The CAC framework conceptualises behaviour as a sequential psychological process in which individuals’ cognitive evaluations influence their affective responses, which subsequ...

work page 2025

[4] [4]

Research Questions and Hypothesis Development 4.1 Enabling Pathway of Willingness to Disclose AI Use Based on the conceptual model, the first research question (RQ1) asks how psychological safety, perceived fairness, perceived teacher support, and perceived organisational support influence university students’ willingness to disclose AI use. Psychological...

work page 2023

[5] [5]

Methods 5.1 Research Design This study adopted a sequential explanatory mixed-methods design, in which quantitative data collection and analysis were followed by qualitative inquiry to further interpret the quantitative findings (Cohen et al., 2018). In the first phase, a questionnaire survey was administered to examine the relationships among the constru...

work page 2018

[6] [6]

Discussion 7.1 Enabling Pathway of Willingness to Disclose AI Use The quantitative findings provide clear support for the enabling pathway proposed in the Cognition–Affect–Conation framework, demonstrating that psychological safety is a central mechanism shaping students’ willingness to disclose AI use. Structural modelling showed that psychological safet...

work page 2024

[7] [7]

The findings show that psychological safety plays a central role in encouraging disclosure, shaped by students’ perceptions of fairness, teacher support, and organisational support

Conclusion This study examined university students’ willingness to disclose AI use through the Cognition–Affect–Conation framework by identifying both enabling and inhibitory psychological pathways. The findings show that psychological safety plays a central role in encouraging disclosure, shaped by students’ perceptions of fairness, teacher support, and ...

work page doi:10.1108/ijis-08-2022-0165 2025

[8] [8]

watching-eye

https://doi.org/10.3390/informatics12020036 Doğan, M., Celik, A., & Arslan, H. (2025). AI in higher education: Risks and opportunities from the academician perspective. European Journal of Education, 60(1), e12863. https://doi.org/10.1111/ejed.12863 Du, C., Tang, M., Wang, C., Zou, B., Xia, Y ., & Du, Y . (2025). Who is most likely to accept AI chatbots? ...

work page doi:10.3390/informatics12020036 2025

[9] [9]

https://doi.org/10.1186/s40468-023-00219-3 Jeilani, A., & Abubakar, S. (2025). Perceived institutional support and its effects on student perceptions of AI learning in higher education: The role of mediating perceived learning outcomes and moderating technology self-efficacy. Frontiers in Education, 10, 1548900. https://doi.org/10.3389/feduc.2025.1548900 ...

work page doi:10.1186/s40468-023-00219-3 2025

[10] [10]

Z., & Belton, B

https://doi.org/10.1186/s40359-025-02412-w Prasad, K., Aladini, A., Normurodova, N. Z., & Belton, B. (2023). Take language assessment easy: The mediator impacts of self-assessment, test-taking skills in predicting student evaluation apprehension, foreign language learning self-esteem, and language achievement in online classes. Language Testing in Asia, 13(1),

work page doi:10.1186/s40359-025-02412-w 2023

[11] [11]

https://doi.org/10.1186/s40468-023-00246-0 Qian, Y . (2025). Pedagogical applications of generative AI in higher education: A systematic review of the field. TechTrends, 69(5), 1105–1120. https://doi.org/10.1007/s11528-025-01100-1 Somerville, S. G., Harrison, N. M., & Lewis, S. A. (2023). Twelve tips for the pre-brief to promote psychological safety in si...

work page doi:10.1186/s40468-023-00246-0 2025

[12] [12]

https://doi.org/10.3390/ijerph19127499 Usher, M., & Barak, M. (2024). Unpacking the role of AI ethics online education for science and engineering students. International Journal of STEM Education, 11(1),

work page doi:10.3390/ijerph19127499 2024

[13] [13]

https://doi.org/10.1186/s40594-024-00493-4 Usman, M., Cheng, J., Ghani, U., Gul, H., & Shah, W. U. (2021). Social support and perceived uncertainties during COVID-19: Consequences for employees’ wellbeing. Current Psychology. https://doi.org/10.1007/s12144-021-02293-3 Vella, S. A., Mayland, E., Schweickle, M. J., Sutcliffe, J. T., McEwan, D., & Swann, C. ...

work page doi:10.1186/s40594-024-00493-4 2021

[14] [14]

https://doi.org/10.3390/jtaer20020099 Zhou, T., & Wu, X. (2025b). Examining generative AI user disclosure intention: An ELM perspective. Universal Access in the Information Society, 24(2), 1209–1220. https://doi.org/10.1007/s10209-024-01130-1 Zhou, T., & Zhang, C. (2024). Examining generative AI user addiction from a C-A-C perspective. Technology in Socie...

work page doi:10.3390/jtaer20020099 2024