Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 79 of the 79 references it cites were found in a registry or an index.

Large language models, social demography, and hegemony: comparing authorship in human and synthetic text

Open access

Published version of a preprint

Abstract

Large language models have become popular over a short period of time because they can generate text that resembles human writing across various domains and tasks. The popularity and breadth of use also put this technology in the position to fundamentally reshape how written language is perceived and evaluated. It is also the case that spoken language has long played a role in maintaining power and hegemony in society, especially through ideas of social identity and “correct” forms of language. But as human communication becomes even more reliant on text and writing, it is important to understand how these processes might shift and who is more likely to see their writing styles reflected back at them through modern AI. We therefore ask the following question: who does generative AI write like? To answer this, we compare writing style features in over 150,000 college admissions essays submitted to a large public university system and an engineering program at an elite private university with a corpus of over 25,000 essays generated with GPT-3.5 and GPT-4 to the same writing prompts. We find that human-authored essays exhibit more variability across various individual writing style features (e.g., verb usage) than AI-generated essays. Overall, we find that the AI-generated essays are most similar to essays authored by students who are males with higher levels of social privilege. These findings demonstrate critical misalignments between human and AI authorship characteristics, which may affect the evaluation of writing and calls for research on control strategies to improve alignment.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s40537-024-00986-7

Cites

79 of 79 references verified0 citation errors · 0 not checkable

Statistics

Figures

Image 1
CC-BY-NC-ND · From the PDF
Fig. 1 Distributions of ( a ) Sixltr (usage of words with six or more letters) and ( b ) affiliation (affiliations) for each set of essays
Fig. 1 Distributions of ( a ) Sixltr (usage of words with six or more letters) and ( b ) affiliation (affiliations) for each set of essaysCC-BY-NC-ND · From the PDF
Fig. 2 Distribution of verb (verb usage) for each set of essays split by gender for ( a ) public school applicants and ( b ) private school applicants
Fig. 2 Distribution of verb (verb usage) for each set of essays split by gender for ( a ) public school applicants and ( b ) private school applicantsCC-BY-NC-ND · From the PDF
Fig. 3 Distributions of analytic (analytical thinking) for each set of essays split by gender for public school applicants ( a ) and private school applicants ( b )
Fig. 3 Distributions of analytic (analytical thinking) for each set of essays split by gender for public school applicants ( a ) and private school applicants ( b )CC-BY-NC-ND · From the PDF
Fig. 4 Distribution verb (usage of common verbs) for each set of essays split by first-generation status for ( a ) public school applicants and ( b ) private school applicants
Fig. 4 Distribution verb (usage of common verbs) for each set of essays split by first-generation status for ( a ) public school applicants and ( b ) private school applicantsCC-BY-NC-ND · From the PDF
Fig. 5 Distribution of Analytic (analytical thinking) for each set of essays split by EC quartiles for ( a ) public school applicants and ( b ) private school applicants
Fig. 5 Distribution of Analytic (analytical thinking) for each set of essays split by EC quartiles for ( a ) public school applicants and ( b ) private school applicantsCC-BY-NC-ND · From the PDF
Fig. 6 Distributions of EC scores at the ZIP code level and predicted values for AI essays with individual-level EC
Fig. 6 Distributions of EC scores at the ZIP code level and predicted values for AI essays with individual-level ECCC-BY-NC-ND · From the PDF
Image 8
CC-BY-NC-ND · From the PDF

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

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?

Self-check

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

The self-check needs the paper’s full text.

Citations

Cites

Show all 72 references

Research works printed but not linked: 20

  1. Bellemare MG, Dabney W, Munos R. A distributional perspective on reinforcement learning. In: International conference on machine learning, PMLR, pp 449–458. 2017.
  2. Bohacek M, Farid H. Nepotistically trained generative-AI models collapse. arXiv preprint. 2023. arXiv:2311.12202. · arXiv
  3. Dhingra H, Jayashanker P, Moghe S, et al. Queer people are people first: deconstructing sexual identity stereotypes in large language models. http://arxiv.org/abs/2307.00101, [cs] 2023. · arXiv
  4. Dominguez-Olmedo R, Hardt M, Mendler-Dünner C. Questioning the survey responses of large language models. arXiv preprint arXiv:2306.07951. 2023. · arXiv
  5. Gallegos IO, Rossi RA, Barrow J, et al. Bias and fairness in large language models: a survey. http://arxiv.org/abs/2309.00770, [cs] 2023. · arXiv
  6. Gramsci A. Prison notebooks, vol. 1. New York: Columbia University Press; 1992.
  7. Kim J, Lee B. Ai-augmented surveys: leveraging large language models for opinion prediction in nationally representative surveys. arXiv preprint arXiv:2305.09620. 2023. · arXiv
  8. Kirk HR, Whitefield A, Röttger P, et al. The PRISM alignment project: what participatory, representative and individualised human feedback reveals about the subjective and multicultural alignment of large language models. arXiv preprint. 2024. arXiv:2404.16019. · arXiv
  9. Lee J, Thymes B, Zhou J, et al. Augmenting holistic review in university admission using natural language processing for essays and recommendation letters. arXiv preprint arXiv:2306.17575. 2023. · arXiv
  10. Lee J, Hicke Y, Yu R, et al (in press) The life cycle of large language models in education: a framework for understanding sources of bias. Br J Educ Technol.
  11. Liang W, Izzo Z, Zhang Y, et al. Monitoring AI-modified content at scale: a case study on the impact of chatgpt on ai conference peer reviews. arXiv preprint. 2024. arXiv:2403.07183. · arXiv
  12. Lucy L, Blodgett SL, Shokouhi M, et al. “One-size-fits-all”? Observations and expectations of NLG systems across identity-related language features. http://arxiv.org/abs/2310.15398, [cs] 2023. · arXiv
  13. Matelsky JK, Parodi F, Liu T, et al. A large language model-assisted education tool to provide feedback on open-ended responses. http://arxiv.org/abs/2308.02439, [cs] 2023. · arXiv
  14. Padmakumar V, He H. Does writing with language models reduce content diversity? In: The Twelfth International Conference on Learning Representations. 2023.
  15. Shen H, Wu T. Parachute: evaluating interactive human-LM co-writing systems. arXiv:2303.06333. [cs] 2023. · arXiv
  16. Stevens ML. Creating a class. Cambridge: Harvard University Press; 2009.
  17. Tao Y, Viberg O, Baker RS, et al. Auditing and mitigating cultural bias in llms. arXiv preprint arXiv:2311.14096. 2023. · arXiv
  18. Wang A, Morgenstern J, Dickerson JP. Large language models cannot replace human participants because they cannot portray identity groups. arXiv preprint arXiv:2402.01908. 2024. · arXiv
  19. Workshop B, Scao TL, Fan A, et al (2022) Bloom: a 176b-parameter open-access multilingual language model. arXiv preprint arXiv:2211.05100. · arXiv
  20. Wu S, Irsoy O, Lu S, et al. BloombergGPT: a large language model for finance. http://arxiv.org/abs/2303.17564, [cs, q-fin] 2023. · arXiv

Cited by

See all 75

Metadata

Authors
Alvero, A. J., Lee, J., Regla-Vargas, A., Kizilcec, R. F., Joachims, T., & Antonio, A. L.
Affiliations
  • Cornell University (US)
  • University of Pennsylvania (US)
  • Stanford University (US)
Corresponding author
  • Alvero, A. J.
  • Antonio, Anthony Lising
Venue
Journal of Big Data
Published
2024-09-27
Volume
11
Issue
1
ISSN
2196-1115 (ISSN-L)
OpenAlex
W4402923080
Cited by
48
References
72
Citations by year
2026: 17 · 2025: 26 · 2024: 4
FWCI
6.94
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal Of Big Data (publishedVersion, cc-by-nc-nd)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • SocArXiv (OSF Preprints) (submittedVersion, cc-by-nd)
Funders
  • Yale University
  • University of Warwick
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Authorship Attribution and Profiling
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
  • Social Sciences › Linguistics and Language › Discourse Analysis in Language Studies
Keywords
large language models, generative AI, authorship, writing style, social demography, hegemony, social privilege, gender differences, college admissions essays, GPT-3.5, GPT-4, writing assessment, human-AI comparison, language and power

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

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Large language models, social demography, and hegemony: comparing authorship in human and synthetic text. https://app.cochat.ai/papers/10.1186/s40537-024-00986-7
The paper
Alvero, A. J., Lee, J., Regla-Vargas, A., Kizilcec, R. F., Joachims, T., & Antonio, A. L. (2024). Large language models, social demography, and hegemony: comparing authorship in human and synthetic text. Journal of Big Data. https://doi.org/10.1186/s40537-024-00986-7

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.