Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

The Digital Divide in Generative AI: Evidence from Large Language Model Use in College Admissions Essays

Jinsook Lee, Conrad Borchers, AJ Alvero, Thorsten Joachims, Rene F. Kizilcec

TL;DR

Concerns about equity and the validity of essay-based evaluation in an era of AI-assisted writing are raised and the first large-scale longitudinal evidence linking LLM adoption, linguistic change, and evaluative outcomes in college admissions is provided.

Abstract

Large language models (LLMs) have become popular writing tools among students and may expand access to high-quality feedback for students with less access to traditional writing support. At the same time, LLMs may standardize student voice or invite overreliance. This study examines how adoption of LLM-assisted writing varies across socioeconomic groups and how it relates to outcomes in a high-stakes context: U.S. college admissions. We analyze a de-identified longitudinal dataset of applications to a selective university from 2020 to 2024 (N = 81,663). Estimating LLM use using a distribution-based detector trained on synthetic and historical essays, we tracked how student writing changed as LLM use proliferated, how adoption differed by socioeconomic status (SES), and whether potential benefits translated equitably into admissions outcomes. Using fee-waiver status as a proxy for SES, we observe post-2023 convergence in surface-level linguistic features, with the largest changes in fee-waived and rejected applicants. Estimated LLM use rose sharply in 2024 across all groups, with disproportionately larger increases among lower SES applicants, consistent with an access hypothesis in which LLMs substitute for scarce writing support. However, increased estimated LLM use was more strongly associated with declines in predicted admission probability for lower SES applicants than for higher SES applicants, even after controlling for academic credentials and stylometric features. These findings raise concerns about equity and the validity of essay-based evaluation in an era of AI-assisted writing and provide the first large-scale longitudinal evidence linking LLM adoption, linguistic change, and evaluative outcomes in college admissions.

The Digital Divide in Generative AI: Evidence from Large Language Model Use in College Admissions Essays

TL;DR

Concerns about equity and the validity of essay-based evaluation in an era of AI-assisted writing are raised and the first large-scale longitudinal evidence linking LLM adoption, linguistic change, and evaluative outcomes in college admissions is provided.

Abstract

Large language models (LLMs) have become popular writing tools among students and may expand access to high-quality feedback for students with less access to traditional writing support. At the same time, LLMs may standardize student voice or invite overreliance. This study examines how adoption of LLM-assisted writing varies across socioeconomic groups and how it relates to outcomes in a high-stakes context: U.S. college admissions. We analyze a de-identified longitudinal dataset of applications to a selective university from 2020 to 2024 (N = 81,663). Estimating LLM use using a distribution-based detector trained on synthetic and historical essays, we tracked how student writing changed as LLM use proliferated, how adoption differed by socioeconomic status (SES), and whether potential benefits translated equitably into admissions outcomes. Using fee-waiver status as a proxy for SES, we observe post-2023 convergence in surface-level linguistic features, with the largest changes in fee-waived and rejected applicants. Estimated LLM use rose sharply in 2024 across all groups, with disproportionately larger increases among lower SES applicants, consistent with an access hypothesis in which LLMs substitute for scarce writing support. However, increased estimated LLM use was more strongly associated with declines in predicted admission probability for lower SES applicants than for higher SES applicants, even after controlling for academic credentials and stylometric features. These findings raise concerns about equity and the validity of essay-based evaluation in an era of AI-assisted writing and provide the first large-scale longitudinal evidence linking LLM adoption, linguistic change, and evaluative outcomes in college admissions.
Paper Structure (45 sections, 4 equations, 7 figures, 13 tables)

This paper contains 45 sections, 4 equations, 7 figures, 13 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. College Admission Essays and Socioeconomic Groups
  4. Risks of LLM-Assisted Writing
  5. Socioeconomic Differences in LLM Adoption and Use
  6. Methodology
  7. Dataset and Preprocessing
  8. LLM Usage Estimation
  9. Synthetic essay generation
  10. Estimation of LLM-generated text
  11. LLM Usage Difference in SES groups in the Post-GPT Period
  12. Measuring Linguistic Changes in Admission Essays
  13. Admission Outcomes and Estimated LLM Usage
  14. Difference-in-Differences Analysis
  15. Stratified Models
  16. ...and 30 more sections

Figures (7)

  • Figure 1: Linguistic changes in college admission essays over time (2020--2024), stratified by socioeconomic status (a) and admission outcome (b). Essays exhibit substantial post-2023 convergence in surface-level linguistic features, with larger changes among lower SES and rejected applicants. Shaded regions indicate 95% confidence intervals; the vertical dashed line marks ChatGPT's public release (Nov.30, 2022).
  • Figure 2: Estimated LLM usage ($\hat{\alpha}$) in admission essays (2020--2024), by socioeconomic status and admission outcome. Usage rises sharply after 2023, with larger increases among lower SES and rejected applicants. Shaded regions denote 95% confidence intervals; the dashed line marks ChatGPT's public release.
  • Figure 3: LLM Usage Distribution by SES Status in the Post-GPT Period.
  • Figure 4: Admission rates by socioeconomic status across pre-GPT and post-GPT eras.
  • Figure 5: Distribution of $\hat{\alpha}$
  • ...and 2 more figures