Cited by 0Doctoral Education Challenges and SolutionsStudent Stress and CopingArtificial Intelligence in Education

The DOI matches its Crossref record; the references it cites are not checked yet; none of the 61 sources its claims cite is checked for support yet.

AI Academic Dependency and Academic Burnout among Doctoral Students: A Cross-Sectional Study of Academic Self-Efficacy, Impostor Feelings, and Fear of Negative Evaluation

Open access
Long, Fei, Xu, Ziqing, Ngah, Abdul Hafaz, Wang, Huilin·2026·International Journal of Mental Health Promotion

Abstract

Background: Generative artificial intelligence (AI) has become increasingly integrated into doctoral students’ academic work, but excessive reliance on these tools may be associated with students’ confidence, perceived academ... | Find, read and cite all the research you need on Tech Science Press

Checked 2026-11-11 · 38 references not checked yet

Overview

AI snapshot

Written from the paper’s full text.

TL;DR

In a cross-sectional survey of 284 doctoral students in China, AI academic dependency was positively associated with academic self-efficacy, impostor feelings, and fear of negative evaluation, and fear of negative evaluation was positively associated with academic burnout. The impostor-related indirect pathways were supported, while the self-efficacy-related indirect effects were statistically significant but operated opposite to the hypothesized directions.

Key findings

  • AI academic dependency was positively associated with academic self-efficacy (β = 0.315, p < 0.001), contrary to H1a, and also with impostor phenomenon (β = 0.325, p < 0.001) and fear of negative evaluation (β = 0.183, p = 0.001).
  • Academic self-efficacy was positively associated with fear of negative evaluation (β = 0.168, p = 0.003), contrary to H2a; impostor phenomenon was positively associated with fear of negative evaluation (β = 0.295, p < 0.001), and fear of negative evaluation was positively associated with academic burnout (β = 0.389, p < 0.001).
  • Significant indirect effects included AD → IP → FN (β = 0.096, 95% CI [0.052, 0.151]), AD → FN → AB (β = 0.071, 95% CI [0.027, 0.124]), and AD → IP → FN → AB (β = 0.037, 95% CI [0.019, 0.065]).
  • Self-efficacy-related indirect effects were significant but not supported as hypothesized: AD → AS → FN (β = 0.053, 95% CI [0.016, 0.101]) and AD → AS → FN → AB (β = 0.021, 95% CI [0.006, 0.044]). R² values were 0.099 for AS, 0.105 for IP, 0.238 for FN, and 0.151 for AB.

Methods

  • Cross-sectional online questionnaire survey of 284 currently enrolled doctoral students in China, recruited via convenience and snowball sampling in May 2026; 300 questionnaires were submitted and 16 excluded for duplicate IPs or uniform response patterns.
  • Measures were adapted to doctoral academic contexts: 6-item AI academic dependency, 8-item academic self-efficacy, 6-item academic impostor phenomenon, 8-item fear of negative evaluation, and 10-item academic burnout, all on 5-point Likert scales.
  • Partial least squares structural equation modelling (PLS-SEM) in SmartPLS 4.1.1.4 was used, with bootstrapping of 5000 resamples to assess direct and indirect effects.

Limitations

  • The paper states that the cross-sectional design prevents causal or temporal inferences, and that self-report questionnaires may introduce common-method and social-desirability bias, especially for sensitive constructs such as AI dependency and impostor feelings.
  • The paper states that convenience and snowball sampling of doctoral students in China limits generalizability, and that the exact number/type of institutions and potential clustering could not be determined because no university, department, or laboratory identifiers were collected.
  • The paper states measurement limitations: the adapted academic burnout scale’s dimensional structure needs further validation in doctoral populations, AI academic dependency was treated as an overall construct without distinguishing forms of AI use, and AI fatigue and changes in AI use were not assessed.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.32604/ijmhp.2026.089822

Cites

0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

47 Sources not found · 34 Own result · 30 Own method · 21 Own interpretation · 7 No basis given

Show all 74 claims

  • It may state two separate claims as one. 28
  • Probably not a substantial claim. 48
65 of 65
  • The paper relies on this without support.

    The deeper psychological mechanisms through which AI dependency may aggravate burnout among doctoral students are substantially under-theorized and empirically neglected.

    “However, the deeper psychological mechanisms through which AI dependency may progressively aggravate burnout among the most academically vulnerable population remain substantially under-theorized and empirically neglected.”
  • The paper relies on this without support.

    The introduction of AI tools capable of performing core academic tasks reveals a novel and largely unexplored risk factor that may intensify existing vulnerabilities among PhD students.

    “Against this already fragile psychological backdrop, the introduction of AI tools capable of performing core academic tasks reveals a novel and largely unexplored risk factor that may intensify existing vulnerabilities among PhD students.”
  • The paper relies on this without support.

    The five constructs (AI academic dependency, academic self-efficacy, impostor phenomenon, fear of negative evaluation, academic burnout) are theoretically distinct aspects of doctoral students' experiences.

    “Although all five constructs were assessed through individual self-reports, they represent theoretically distinct aspects of doctoral students' experiences.”
  • The paper relies on this without support.

    The cross-sectional associations reflect the hypothesized temporal ordering, with AI academic dependency preceding the mediators and outcomes.

    “This study adopted a cross-sectional online questionnaire design”
    “The indirect effect of AI academic dependency on fear of negative evaluation through academic self-efficacy was positive and significant”
  • The paper relies on this without support.

    The convenience and snowball sample of 284 doctoral students in China is representative enough to generalize the findings to doctoral students in China.

    “employed convenience and snowball sampling to recruit currently enrolled doctoral students in China”
    “A total of 300 questionnaires were submitted.Responses were screened using predefined quality-control criteria.Sixteen responses were excluded because of duplicate IP addresses or highly uniform response patterns, leaving 284 participants in the final analytical sample.”
  • The paper relies on this without support.

    The adapted measures adequately capture the intended constructs in the doctoral academic context.

    “The items were selected on the basis of conceptual relevance to the present academic context rather than their factor loadings in the original validation study.”
    “The adapted six-item measure demonstrated good internal consistency in the present sample (Cronbach's α = 0.897).”
  • The paper relies on this without support.

    Academic environments may differ across countries in supervisory practices, competitive norms, performance expectations, evaluation systems, and attitudes toward generative AI.

    “Academic environments may differ across countries in supervisory practices, competitive norms, performance expectations, evaluation systems, and attitudes toward generative AI.”
  • The paper relies on this without support.

    Undergraduate and master's students may experience different academic demands, evaluative contexts, and patterns of generative AI use.

    “Undergraduate and master's students may also experience different academic demands, evaluative contexts, and patterns of generative AI use.”
8 of 8

Citations

Cites

The publisher hasn’t deposited this paper’s reference list.

Research works printed but not linked: 8

  1. Zhao T, Zhao X, Cheah KSL, Zhang Y. ChatGPT, loneliness, and well-being among international PhD students in Malaysia: a mixed-methods study. Int J Ment Health Promot. 2025;27(12):2023-38. [CrossRef]. · DOI
  2. Naseer A, Ahmad NR, Chishti MA. Psychological impacts of AI dependence: assessing the cognitive and emotional costs of intelligent systems in daily life. Rev Appl Manag Soc Sci. 2025;8(1):291-307. [CrossRef]. · DOI
  3. Guo Y, Liu X. Mental health and well-being of doctoral students: a systematic review. Int J Ment Health Promot. 2026;28(1):1-10. [CrossRef]. · DOI
  4. Schunk DH, DiBenedetto MK. Academic self-efficacy. In: Handbook of positive psychology in schools: supporting process and practice. New York, NY, USA: Routledge; 2022. p. 268-82.
  5. Pajares F. Self-efficacy beliefs in academic settings. Rev Educ Res. 1996;66(4):543-78. [CrossRef]. · DOI · Paper page
  6. de la Fuente J, Kauffman DF, Boruchovitch E. Editorial: past, present and future contributions from the social cognitive theory (Albert Bandura). Front Psychol. 2023;14:1258249. [CrossRef]. · DOI
  7. Fariborz N, Hadi J, Ali TN. Students' academic stress, stress response and academic burnout: mediating role of self-efficacy. Pertanika J Soc Sci Humanit. 2019;27(4):2441-54.
  8. Bandura A. The role of selective moral disengagement in terrorism and counterterrorism. In: Understanding terrorism: Psychosocial roots, consequences, and interventions. Washington, DC, USA: American Psychological Association; 2005. p. 121-50. [CrossRef]. · DOI

Cited by

Still being checked.

Metadata

Authors
Long, F., Xu, Z., Ngah, A. H., & Wang, H.
Affiliations
  • Guangdong Ocean University (CN)
  • Universiti Malaysia Terengganu (MY)
  • Hunan University of Science and Technology (CN)
Corresponding author
Xu, Ziqing
Venue
International Journal of Mental Health Promotion
Published
2026
Volume
0
Issue
0
Pages
1-10
ISSN
  • 1462-3730 (ISSN-L)
  • 2049-8543
OpenAlex
W7225122516
Cited by
0
References
0
FWCI
0.00
Type
journal article
Language
en
Open access
Locations
International Journal of Mental Health Promotion (publishedVersion)
Funders
Guangdong Ocean University
Topics
  • Social Sciences › Education › Doctoral Education Challenges and Solutions
  • Social Sciences › Psychology › Developmental and Educational Psychology › Student Stress and Coping
  • Social Sciences › Education › Artificial Intelligence in Education
Keywords
generative AI, doctoral students, academic burnout, academic self-efficacy, fear of negative evaluation, AI reliance, academic mental health

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review