AI May Broaden Writing Help While Deepening an Admissions Divide

The finding

A study of 81,663 applications to a selective university found rapid post-2023 changes in essay language across socioeconomic groups, with especially large changes among fee-waiver applicants and students who were ultimately rejected. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays.

A large longitudinal study suggests lower-income applicants adopted AI-assisted writing quickly—but did not receive equal admissions benefit from the shift. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays.

What the evidence actually shows

Researchers used fee-waiver status as a socioeconomic proxy and a distribution-based detector to estimate changes associated with large-language-model writing. They report that estimated AI use rose across groups, but the relationship between estimated use and predicted admission probability was more negative for lower-SES applicants after controlling for academic credentials and writing features. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays. Common App: Fee waiver guidance.

The decision hidden inside the headline

The access story is double-edged. Low-cost AI can supply feedback to students who cannot buy private editing, yet the resulting prose may converge toward patterns that readers distrust or that obscure the lived detail selective review is supposed to surface. Equal access to a tool does not guarantee equal benefit when applicants enter with different counseling, time and revision support. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays. Common App: Fee waiver guidance.

What the record cannot yet prove

The detector estimates patterns rather than observing every student’s actual writing process, and fee-waiver status is an incomplete measure of socioeconomic background. The work identifies an equity concern, not proof that a college intentionally penalized a protected group. Common App: Fee waiver guidance.

What applicants should do now

Schools and advisers should teach transparent, limited AI use alongside human revision. Applicants using fee waivers should not be pushed toward generic automated prose; they need specific feedback that protects their own voice, context and factual accuracy. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays.

Preserve the exact cycle, program and applicant category in any checklist. When a secondary report and a controlling admissions page differ, the current first-party instruction governs the application. arXiv: The Digital Divide in Generative AI: Evidence from College Admissions Essays.