CoChat

A Comparative Analysis of AI Privacy Concerns in Higher Education: News Coverage in China and Western Countries

Open access
Xue, Yujie, Chinapah, Vinayagum, Zhu, Chang · 2025 · Education Sciences

The DOI matches its Crossref record; 41 of the 41 references it cites were found in a registry or an index; 11 of the 28 sources its claims cite were checked for support.

Abstract

This study examines how Chinese and Western news media covered artificial intelligence (AI) privacy issues in higher education from 2019 to 2024. News articles were retrieved from Nexis Uni. First, non-negative matrix factorization (NMF) was employed to identify core AI privacy topics in university teaching, administration, and research. Next, a time trend analysis investigated how media attention shifted in relation to key events, including the COVID-19 pandemic and the emergence of generative AI. Finally, a sentiment analysis was conducted to compare the distribution of positive, negative, and neutral reporting. The findings indicate that AI-driven proctoring, student data security, and institutional governance are central concerns in both Chinese and English media. However, the focus and framing differ: some Western outlets highlight individual privacy rights and controversies in remote exam monitoring, while Chinese coverage more frequently addresses AI-driven educational innovation and policy support. The shift to remote education after 2020 and the rise of generative AI from 2023 onward have intensified discussions on AI privacy in higher education. The results offer a cross-cultural perspective for institutions seeking to reconcile the adoption of AI with robust privacy safeguards and provide a foundation for future data governance frameworks under diverse regulatory environments.

Overview

CoChat Summary

TL;DR

This study compares Chinese and Western news coverage of AI privacy in higher education from 2019 to 2024 using NMF topic modeling, time trend analysis, and lexicon-based sentiment analysis on 315 English and 347 Chinese articles. It finds shared concerns with AI proctoring, student data security, and institutional governance, but Western coverage is more event-driven and critical/negative while Chinese coverage is more policy-driven and positive.

Key findings

  • NMF identified four topics in each corpus. Western media topics were AI and Big Tech in University Privacy Regulation; AI in Higher Education and Student Privacy; AI Surveillance and Facial Recognition in Universities; and AI Research, Geopolitics, and University Collaborations. Chinese media topics were Digital Industry and Cybersecurity; Education and Personalized Learning; Governance and ethical supervision; and AI development and application.
  • Chinese media showed overwhelming positive sentiment on AI privacy, with more than 80% of news positive; Western media had a more diversified distribution with a much larger proportion of negative and neutral articles. Topic-specific sentiment: Chinese Topic 1 (digital industry/cybersecurity) and Topic 3 (governance/ethical supervision) were 93% and 83% positive, while Chinese Topic 2 (education/personalized learning) had 26% negative sentiment; Western negative reports accounted for 69% for Topic 1, 61% for Topic 3, and 35% for Topic 2.
  • Both media increased attention after 2022, but Western coverage was intermittent and event-driven, tied to GDPR enforcement, student protests against biometric exams, and China-US academic cooperation reviews; Chinese coverage showed policy-driven systematic growth after late 2022, peaking in 2024, linked to PIPL and generative AI governance.
  • Common focuses included AI-driven proctoring, student data security, and institutional governance. Differences: Western media emphasized individual privacy rights, legal accountability, and tech company regulation under GDPR/FERPA, while Chinese media emphasized AI-driven educational innovation, policy support, and government-led governance under PIPL.

Methods

  • Data came from Nexis Uni for 1 January 2019 to 31 December 2024. The English search yielded 1574 results from The New York Times, The Guardian, Times Higher Education, and The Chronicle of Higher Education; after manual screening, 315 articles were retained. The Chinese initial search returned 839 entries; after filtering, 347 valid entries were retained from China Industry Information Database, Xinhua News Agency: Overseas Service Chinese News, Beijing Review, and Xinhua Finance and Economics Chinese.
  • Analysis used NMF topic modeling on TF-IDF matrices, with k = 4 selected after testing k values from three to six based on semantic interpretability; time trend analysis by publication date; and lexicon-based sentiment analysis with custom Chinese and English lexicons, Jieba segmentation for Chinese, and negation handling. No direct interaction with human subjects was involved.

Limitations

  • Stated: the data source is limited to news media and may not reflect direct views of university administrators, students, and faculty; sentiment tools may struggle with implicit emotions, sarcasm, or culturally specific affective expressions; the number of topics (k) and sentiment labeling were based on manual evaluation rather than automated validation, and no formal coherence scoring was applied.
  • Stated/evident: Chinese sources are primarily state-affiliated and operate in a centralized media environment, which may influence framing; the Western corpus includes a wider range of commercial and independent outlets; the study does not perform a detailed source-level comparison, and further analysis of source diversity—particularly within Chinese media—remains a future direction.

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

Integrity signals

Registry

Crossref · 10.3390/educsci15060650

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

11 Valid · 6 Not valid · 17 Not checked · 112 Own results and interpretations

57 more claims

Create a free account to see all 146 claims and to verify the 3 references still open.

See all claims — free account
  • It may state two separate claims as one. 57
  • Probably not a substantial claim. 38

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Cited by

See all 30

Metadata

Authors
Xue, Y., Chinapah, V., & Zhu, C.
Affiliations
  • Vrije Universiteit Brussel (BE)
  • Stockholm University (SE)
Corresponding author
Zhu, Chang
Venue
Education Sciences
Publisher
MDPI AG
Published
2025-05-25
Volume
15
Issue
6
Pages
650
ISSN
2227-7102 (ISSN-L)
OpenAlex
W4410721456
Cited by
30
References
36
Citations by year
2026: 22 · 2025: 8
FWCI
31.45
Type
journal article
Language
en
Open access
Locations
  • Education Sciences (publishedVersion, cc-by)
  • VUBIR (Vrije Universiteit Brussel) (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Sociology and Political Science › Privacy, Security, and Data Protection
Keywords
artificial intelligence in higher education, AI privacy, news media coverage, media framing, generative AI, student data security, institutional governance, COVID-19 pandemic, cross-cultural media comparison, China, Western countries, sentiment analysis, data governance

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.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

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

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.