pith. sign in

arxiv: 2605.16642 · v1 · pith:V5NMIAGJnew · submitted 2026-05-15 · 💻 cs.HC

Examining University Students' Artificial Intelligence-Generated Content (AIGC) Verification Intention from a Protection Motivation Perspective

Yiran Du This is my paper

Pith reviewed 2026-05-20 15:40 UTC · model grok-4.3

classification 💻 cs.HC
keywords AIGC verificationProtection Motivation Theoryuniversity studentsartificial intelligenceSEMfsQCAthreat appraisalcoping appraisal
0
0 comments X

The pith

Protection motivation from recognizing AI content risks and coping ability drives students' intention to verify AIGC outputs.

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

This paper applies Protection Motivation Theory to university students' decisions about checking AI-generated content for learning tasks. It tests how threat appraisal factors such as perceived severity of inaccuracies and personal vulnerability increase protection motivation, while coping appraisal elements like response efficacy and self-efficacy also strengthen it. Maladaptive rewards from skipping verification and response costs reduce motivation. The survey of 432 students with AIGC experience uses structural equation modeling to confirm that higher protection motivation predicts greater verification intention, with fuzzy-set qualitative comparative analysis identifying multiple pathways where protection motivation serves as a core condition. A sympathetic reader would care because AI tools are now common in higher education yet often produce errors, and identifying what prompts careful checking could support more responsible academic use.

Core claim

The study finds that protection motivation positively predicts AIGC verification intention. Perceived severity, perceived vulnerability, response efficacy, and self-efficacy positively influence protection motivation, whereas maladaptive rewards and response cost exert negative effects. The fsQCA analysis identifies three configurations leading to high verification intention, with protection motivation appearing as a core condition in all of them.

What carries the argument

Protection Motivation Theory applied to AIGC verification, where threat appraisal and coping appraisal shape protection motivation that then leads to verification intention.

If this is right

  • Students who see greater severity in AIGC inaccuracies will show stronger verification intention through elevated protection motivation.
  • Belief in one's ability to respond effectively to AIGC issues increases protection motivation and subsequent verification intention.
  • High response costs or rewards for not verifying lower protection motivation and reduce verification intention.
  • Multiple combinations of factors can lead to high verification intention as long as protection motivation remains high.

Where Pith is reading between the lines

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

  • Educators could design interventions that heighten awareness of AIGC risks while teaching practical verification skills to boost overall checking rates.
  • The findings might extend to other AI-assisted tasks beyond learning, such as professional writing, if similar threat and coping appraisals apply.
  • Future work could test whether these motivation patterns change as AIGC tools improve in accuracy over time.

Load-bearing premise

The study assumes that self-reported survey measures of threat appraisal, coping appraisal, and protection motivation accurately capture students' real-world verification behaviors without substantial social desirability bias, measurement error, or common method variance.

What would settle it

A follow-up study that tracks actual verification actions during real AIGC use and finds no link to the protection motivation scores measured in the original survey would undermine the claim.

Figures

Figures reproduced from arXiv: 2605.16642 by Yiran Du.

Figure 1
Figure 1. Figure 1: The Conceptual Model 3.1 Protection Motivation Protection motivation refers to an individual’s intention to adopt protective actions when facing a perceived threat (Hinssen & Dohle, 2023). According to Protection Motivation Theory (PMT), protection motivation is formed through threat appraisal and coping appraisal, and it functions as the proximal psychological driver of protective behaviour (Floyd et al.,… view at source ↗
read the original abstract

Artificial Intelligence-Generated Content (AIGC) is increasingly used by students to support learning tasks, yet its outputs may contain inaccuracies, fabricated references, bias, and unsupported claims. This study examined students' intention to verify AIGC from the perspective of Protection Motivation Theory. A cross-sectional survey was conducted with 432 students who had experience using AIGC for learning. Structural equation modelling (SEM) was used to test the hypothesised relationships among threat appraisal, coping appraisal, protection motivation, and AIGC verification intention, while fuzzy-set qualitative comparative analysis (fsQCA) was applied to identify configurational pathways leading to high verification intention. The SEM results showed that protection motivation positively predicted AIGC verification intention. Perceived severity, perceived vulnerability, response efficacy, and self-efficacy positively influenced protection motivation, whereas maladaptive rewards and response cost had negative effects. The fsQCA results further revealed three configurations leading to high verification intention, with protection motivation appearing as a core condition across all pathways. These findings suggest that students' willingness to verify AIGC depends on both risk recognition and perceived coping capacity. The study extends Protection Motivation Theory to the context of AIGC verification and provides implications for promoting critical, responsible, and academically appropriate use of generative AI 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.

Referee Report

2 major / 2 minor

Summary. The manuscript applies Protection Motivation Theory to examine university students' AIGC verification intentions. A cross-sectional survey of 432 students with AIGC experience is analyzed via SEM to test how threat appraisal (severity, vulnerability) and coping appraisal (response efficacy, self-efficacy, maladaptive rewards, response cost) shape protection motivation, which in turn predicts verification intention. fsQCA identifies three configurational pathways to high verification intention, with protection motivation as a core condition in all. The abstract reports directional effects but omits model fit, reliability, and missing-data details.

Significance. If the results hold after methodological strengthening, the work usefully extends PMT to AIGC verification in higher education and demonstrates the value of combining SEM with fsQCA for both net-effect and configurational insights. The findings could inform interventions that promote critical evaluation of generative-AI outputs, addressing documented risks such as inaccuracies and bias.

major comments (2)
  1. [Methods] Methods section: The single-time-point self-report survey design leaves the SEM results vulnerable to common-method variance. No procedural remedies (temporal separation, marker variables) or post-hoc checks (Harman’s test, latent method factor, or CFA with method factor) are reported, directly threatening the validity of the central claim that protection motivation positively predicts verification intention.
  2. [Results] Results section: The manuscript provides no SEM model-fit indices (CFI, RMSEA, SRMR), reliability/validity statistics (Cronbach’s α, CR, AVE), or description of missing-data handling. These omissions are load-bearing because they prevent verification that the reported path coefficients and fsQCA configurations are supported by the data.
minor comments (2)
  1. [Abstract] Abstract: The abstract should include sample size, at least one key fit index, and a brief description of the three fsQCA configurations to give readers a self-contained summary.
  2. [Discussion] Discussion: The practical implications for higher education could be made more concrete by linking specific configurations to actionable recommendations for instructors or platform designers.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. The comments highlight important issues of methodological transparency and reporting that we have addressed in the revision. We provide point-by-point responses below.

read point-by-point responses

  1. Referee: [Methods] Methods section: The single-time-point self-report survey design leaves the SEM results vulnerable to common-method variance. No procedural remedies (temporal separation, marker variables) or post-hoc checks (Harman’s test, latent method factor, or CFA with method factor) are reported, directly threatening the validity of the central claim that protection motivation positively predicts verification intention.

    Authors: We agree that common-method variance is a valid concern for cross-sectional self-report data and that its absence from the original submission is a limitation. In the revised manuscript we will add Harman's single-factor test as a post-hoc check, report the percentage of variance explained by the first factor, and discuss the implications for the protection motivation–verification intention path. We will also note this as a study limitation and suggest future research employ temporal separation or marker variables. revision: yes

  2. Referee: [Results] Results section: The manuscript provides no SEM model-fit indices (CFI, RMSEA, SRMR), reliability/validity statistics (Cronbach’s α, CR, AVE), or description of missing-data handling. These omissions are load-bearing because they prevent verification that the reported path coefficients and fsQCA configurations are supported by the data.

    Authors: We acknowledge these reporting omissions. The revised manuscript will include a table or subsection with SEM fit indices (CFI, RMSEA, SRMR), reliability (Cronbach’s α) and validity (CR, AVE) statistics for all constructs, and a description of missing-data handling (listwise deletion after confirming low missingness). These details were computed during analysis but not fully presented; their inclusion will allow readers to evaluate the robustness of both the SEM paths and the fsQCA solutions. revision: yes

Circularity Check

0 steps flagged

No significant circularity: empirical test of external theory on survey data

full rationale

The paper applies the established Protection Motivation Theory (external to this work) to AIGC verification via standard SEM hypothesis testing and fsQCA on a cross-sectional survey of 432 students. All central claims are empirical associations derived from the collected data against pre-specified hypotheses; no step reduces a prediction to a fitted parameter from the same dataset by construction, nor relies on self-citation chains, ansatzes smuggled via prior work, or renaming of known results. The derivation chain is self-contained against external benchmarks and does not exhibit any of the enumerated circularity patterns.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the validity of Protection Motivation Theory constructs as measured by the survey instrument and on standard statistical assumptions of SEM (linearity, no severe multicollinearity) and fsQCA (appropriate calibration of fuzzy sets).

axioms (2)
  • domain assumption Protection Motivation Theory constructs (threat appraisal, coping appraisal) can be reliably measured via self-report Likert scales in the AIGC context.
    Invoked implicitly by applying the theory to new survey data without additional validation steps mentioned in the abstract.
  • standard math Structural equation modelling assumptions hold for the collected data (multivariate normality, adequate sample size for the model).
    Required for interpreting the reported positive and negative influences on protection motivation.

pith-pipeline@v0.9.0 · 5760 in / 1373 out tokens · 34638 ms · 2026-05-20T15:40:37.140216+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

15 extracted references · 15 canonical work pages · 1 internal anchor

  1. [1]

    Introduction Artificial Intelligence-Generated Content (AIGC) is increasingly embedded in higher education, where it supports students in information retrieval, writing, translation, summarisation, brainstorming, and problem-solving (Yan & Qianjun, 2025; Wang & Zhang, 2025). As generative AI systems become more widely used in learning contexts, AIGC has s...

    work page 2025

  2. [2]

    AIGC generally refers to content produced by generative AI systems, including text, images, code, audio, video, and multimodal outputs (Rizun et al., 2026)

    Literature Review 2.1 Artificial Intelligence-Generated Content Artificial Intelligence-Generated Content (AIGC) has attracted increasing scholarly attention as generative artificial intelligence becomes widely integrated into educational settings (Wu et al., 2026). AIGC generally refers to content produced by generative AI systems, including text, images...

    work page 2026

  3. [3]

    In educational contexts, students’ verification intention is closely related to critical thinking, information literacy, and academic integrity (Sun & Zhou, 2024). Although AIGC can reduce effort and increase productivity, relying on unverified content may result in the reproduction of inaccurate knowledge and inappropriate academic practices (Stone, 2024...

    work page 2024

  4. [4]

    The Conceptual Model 3.1 Protection Motivation Protection motivation refers to an individual’s intention to adopt protective actions when facing a perceived threat (Hinssen & Dohle, 2023). According to Protection Motivation Theory (PMT), protection motivation is formed through threat appraisal and coping appraisal, and it functions as the proximal psychol...

    work page 2023

  5. [5]

    Self-reported questionnaire data were collected from students who had experience using AIGC in learning contexts

    Methodology 4.1 Research Design This study adopted a cross-sectional survey design to examine students’ intention to verify artificial intelligence-generated content (AIGC) from the perspective of Protection Motivation Theory (PMT). Self-reported questionnaire data were collected from students who had experience using AIGC in learning contexts. The resear...

    work page 2023

  6. [6]

    Du, 2024)

    Characteristic Category n % Gender Male 194 44.9 Female 238 55.1 Age 18–20 156 36.1 21–23 201 46.5 24 or above 75 17.4 Study level Undergraduate 287 66.4 Postgraduate 145 33.6 Academic discipline STEM 213 49.3 Non-STEM 219 50.7 4.3 Measurement All constructs were measured using multi-item scales adapted from prior research on Protection Motivation Theory ...

    work page 2024

  7. [7]

    Responses were collected using a five-point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree, with higher scores indicating stronger agreement with the corresponding construct. The original English items were translated into Chinese and then back-translated into English by bilingual researchers to ensure linguistic accuracy and concep...

    work page 2023

  8. [8]

    Because all data were collected through a self-report questionnaire, common method variance (CMV) was assessed (Podsakoff et al., 2024)

    to assess item clarity, wording, and questionnaire flow, and minor revisions were made based on their feedback. Because all data were collected through a self-report questionnaire, common method variance (CMV) was assessed (Podsakoff et al., 2024). Harman’s single-factor test showed that the first unrotated factor accounted for less than 50% of the total ...

    work page 2024

  9. [9]

    Discussion 6.1 Net Effects on AIGC Verification Intention The SEM results indicate that protection motivation is a significant direct predictor of students’ AIGC verification intention. This finding supports the core assumption of Protection Motivation Theory that protective behavioural intention is shaped by individuals’ motivation to avoid or reduce per...

    work page 1975

  10. [10]

    Conclusion This study examined students’ AIGC verification intention from the perspective of Protection Motivation Theory. The SEM results showed that protection motivation directly promoted verification intention, while perceived severity, perceived vulnerability, response efficacy, and self-efficacy strengthened protection motivation, and maladaptive re...

    work page doi:10.1186/s41239-024-00444-7 2024

  11. [11]

    https://doi.org/10.3389/fpsyg.2025.1644209 Balla, J., & Hagger, M. S. (2024). Protection motivation theory and health behaviour: Conceptual review, discussion of limitations, and recommendations for best practice and future research. Health Psychology Review, 19(1), 145–171. https://doi.org/10.1080/17437199.2024.2413011 Basarmak, U., & Ates, H. (2026). In...

    work page doi:10.3389/fpsyg.2025.1644209 2025

  12. [12]

    https://doi.org/10.3389/fpsyg.2023.1195607 Huang, Z., Fu, X., & Zhao, J. (2025). Research on AIGC-integrated design education for sustainable teaching: An empirical analysis based on the TAM and TPACK models. Sustainability, 17(12),

    work page doi:10.3389/fpsyg.2023.1195607 2023

  13. [13]

    https://doi.org/10.3390/su17125497 Ifinedo, P. (2012). Understanding information systems security policy compliance: An integration of the theory of planned behavior and the protection motivation theory. Computers & Security, 31(1), 83–95. https://doi.org/10.1016/j.cose.2011.10.007 Janmaimool, P. (2017). Application of protection motivation theory to inve...

    work page doi:10.3390/su17125497 2012

  14. [14]

    https://doi.org/10.3390/su9071079 Jia, K., Cui, K., He, H., & Du, Y . (2026). The impact of coercive, normative, and mimetic Stress on Chinese teachers’ continuance intention to use generative AI: An integrated perspective of the Expectation-Confirmation Model and Institutional Theory (arXiv:2605.00522). arXiv. https://doi.org/10.48550/arXiv.2605.00522 Ka...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.3390/su9071079 2026

  15. [15]

    https://doi.org/10.1016/j.ijer.2026.102963 Tsai, H

    International Journal of Educational Research, 137, 102963. https://doi.org/10.1016/j.ijer.2026.102963 Tsai, H. S., Jiang, M., Alhabash, S., LaRose, R., Rifon, N. J., & Cotten, S. R. (2016). Understanding online safety behaviors: A protection motivation theory perspective. Computers & Security, 59, 138–150. https://doi.org/10.1016/j.cose.2016.02.009 Wang,...

    work page doi:10.1016/j.ijer.2026.102963 2026