Cited by 16Multilingual Education and PolicyEFL/ESL Teaching and LearningDiscourse Analysis in Language Studies

You don’t need to prove yourself: A raciolinguistic perspective on Chinese international students’ academic language anxiety and ChatGPT use

Open access
Zheng, Kewen·2025·Linguistics and Education

Abstract

• ChatGPT exacerbates language anxiety among Chinese international students. • Further reliance occurs due to ChatGPT’s response mode and readers’ biased policing. • Raciolinguistic ideologies work with negative figures of personhood in AI policing. • Internalized academic English norms engage racialized students in self-policying. Adopting a raciolinguistic perspective, this study examines the Chinese international students’ anxiety on academic English writing under ChatGPT use and AI policing. As the baseline of language proficiency and a gatekeeper to academic achievement, academic English stands as the exclusive linguistic standard within academia. Chinese international students in the U.S. often experience different levels of anxiety in academic English writing as they are usually perceived as outsiders and linguistically deficient. In this context, AI language refining tools like ChatGPT have become a popular tool for such students to “standardize” their academic English writing. While these tools “improve” the quality of writing to some extent, they in fact exacerbate linguistic insecurity and language anxiety, because of their native-like response mode and readers’ biased policing. Through analysis of the interview data and human-AI interaction data provided by 18 Chinese international students in a U.S. graduate program, this study reveals the structural challenges faced by bilingual international students as they navigate academia in the era of ChatGPT. This paper contributes to a deeper understanding of the complexities surrounding requirement of academic English proficiency and racialized speakers’ experiences of being subjugated and engaging in self-policing in academic settings.

Integrity signals

Registry

Crossref · 10.1016/j.linged.2025.101406

Cites

89 of 89 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

We couldn’t retrieve this paper’s full text, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • In English-dominant academic and corporate cultures where grammar is policed and accented English is disparaged, we have come to associate expertise with normative grammar, idiomatic syntax and vocabulary and slick delivery.

    “In English-dominant academic and corporate cultures where grammar is policed and accented English is implicitly or explicitly disparaged (Zheng 2025), we have come to associate expertise with normative grammar, idiomatic syntax and vocabulary and slick delivery.”

    From Large language models and the problem of rhetorical debt (2025) · cites this paper as Zheng 2025

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Zheng, K.
Venue
Linguistics and Education
Publisher
Elsevier BV
Published
2025-04
Volume
86
Pages
101406
ISSN
0898-5898 (ISSN-L)
OpenAlex
W4408056799
Cited by
16
References
88
Citations by year
2026: 12 · 2025: 4
FWCI
53.80
Type
journal article
Language
en
Open access
Locations
Linguistics and Education (publishedVersion, cc-by-nc-nd)
Funders
National University of Singapore
Topics
  • Social Sciences › Linguistics and Language › Multilingual Education and Policy
  • Social Sciences › Arts and Humanities › Language and Linguistics › EFL/ESL Teaching and Learning
  • Social Sciences › Arts and Humanities › Literature and Literary Theory › Discourse Analysis in Language Studies
Keywords
raciolinguistics, Chinese international students, ChatGPT, English academic writing, linguistic insecurity, self-policing, norm internalization, racialized students, bilingual students, human-AI interaction

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.

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